RU2598294C2 - Broadcasting system for alerting mobile communication subscribers on occurrence of emergency situations, user communication device and method of its operation - Google Patents

Abstract

FIELD: radio engineering and communications.

SUBSTANCE: invention relates to systems for mobile wireless communication, more specifically to broadcasting system notification of mobile communication subscribers on occurrence of emergency situations using short text messages, also refers to devices which are included into described system and methods of their functioning. System has single set of telephone numbers. System also comprises facilities, providing automatic recharge of mobile subscribers accounts, information receiving module for actuation of information from means of tracking the situation to single format and dimensions, information transmission module for generation of identical short text messages containing notification situation on limited version of natural language, separation of text into segments, limiting messages length, information transmission module containing text segments data bank, multilayer network specially trained artificial neural network (INS) and its training and configuration means, multiple outputs of which correspond to plurality of text segments.

EFFECT: wider field of use, protection of wireless communication networks against overload.

7 cl, 15 dwg, 3 tbl

Description

FIELD OF THE INVENTION

The invention relates to the field of mobile wireless communication systems, and more specifically to a broadcast system for notifying mobile subscribers about emergencies using short text messages, devices included in the described system, and methods of its operation.

BACKGROUND AND BACKGROUND OF THE INVENTION

Mobile wireless networks are widely deployed in the Russian Federation and abroad, GSM (Global System for Mobile Communications) networks and their modifications are widespread: GPRS (General Packet Radio Service), EDGE (Enhanced Data rates for GSM Evolution), 3G (3 -rd generation GSM), GSM-R (GSM-Railway), etc.

The term “subscriber device” in the future will mean any mobile telephone device capable of operating in such wireless communication networks and having the ability to receive incoming short text messages, including a mobile cell phone, smartphone and tablet computer. At the same time, the “mobile subscriber” (“subscriber”) is a regular user of such a device.

By “short text messages” (“messages”) in this invention, basically, we mean messages of the Short Messaging Service (SMS), which allows you to send and receive alphanumeric messages on the display of a subscriber device, as well as its more advanced modifications, including EMS (Enhanced Messaging Service), USSD (Unstructured Supplementary Service Data), CB (Cell Broadcast), etc.

“Broadcast” means the simultaneous distribution of the same short text messages among multiple subscriber devices.

“Emergency” means a situation that suddenly arises requiring a mobile subscriber to take some urgent measures (take actions). This term includes (but is not limited to only) situations related to the threat to the integrity of property, life and health caused by natural disasters of natural, man-made and / or social origin.

“Emergency notification” (“notification”) is the notification to the subscriber of an emergency, information about the emergency and the recommended procedure for the subscriber.

Recently, smartphones have become widespread among subscriber devices. They, in particular, differ from ordinary mobile phones in the presence of a developed operating system open for the rapid development of additional software. Installing additional programs can improve the functionality of smartphones. The most popular smartphone operating systems today are Google Android, Apple iOS, Microsoft Windows Phone. The number of programs and devices connected to them offered for installation on smartphones is growing, and the range of tasks they are solving is expanding.

At the same time, a significant number of subscribers continue to use “old-style” subscriber devices designed only for voice calls and sending / receiving SMS messages that do not support the ability to install additional programs and connect third-party devices. Often, such subscribers belong to the most vulnerable social groups: elderly people, children, the poor, homeless, etc. For the complete replacement of all subscriber devices of the "old model" with modern smartphones, it will take many years.

The prior art [1] known technology for broadcasting SMS-messages Cell Broadcast (CB), operating in GSM networks and the like. Cell Broadcast is designed to quickly deliver any message in a specific geographic area. This technology was first demonstrated in Paris in 1997.

Cell Broadcast is suitable for alerting the public about emergencies. This is primarily due to the ability to choose the territory of the distribution of messages. In addition, CB messages create minimal additional network load.

To implement the technology in the GSM network, a new Cell Broadcast Center (SHS) element is installed - the broadcast center. It stores information about the location of existing base stations and their geographical location. SHS is connected to base station controllers BSC (Base Station Controller). The introduction of Cell Broadcast technology does not affect the switching system.

If it becomes necessary to send a message to a specific geographical area, the SHS checks the information it has about the location of existing base stations and sends a message only to the necessary base station controllers (BSC), which, in turn, already redirect this message to right honeycombs.

Moreover, not every message goes to each controller, but only one, which is then duplicated at the base stations, which in turn send them to the cells. Thus, the minimum load of network resources and targeted mailing is ensured.

It is known from the prior art [2] that attempts are being made in the world to use mobile wireless communication systems to notify subscribers about emergencies of a different nature using short messages. The systems mentioned below can be considered as analogues of the claimed invention.

For example, the Firewatch warning system via SMS “Firewatch” is used in Australia [3, 4]. Subscribers can subscribe to this service and in case of a fire within a radius of 15 km receive notifications on their subscriber devices.

The system contains sensors mounted on towers in the forest. If the sensor detects a fire, it transmits information about it to the control center via wireless communication. At the control center is personnel who receive and process the received message.

The staff decides on the transfer of information to the management workplace. The head of the control center decides to inform the fire brigades and the population within a radius of 15 km from the place of ignition.

However, warning systems using SMS messages have a number of common disadvantages that significantly reduce the effectiveness of their use:

- the maximum length of an SMS message is limited to 160 characters (when using the Latin alphabet), which is not enough to clearly describe the situation and the procedure;

- a lot of calls, usually made by subscribers in case of an emergency, so overloads the network that it leads to an unacceptably long delay in transmitting and receiving SMS messages;

- using SMS-messages it is possible to receive only alphanumeric messages, while it is necessary to obtain information in a more visual form: graphic (for example, a map with the location of evacuation places), sound (for example, voice message according to the procedure) and / or other species.

- at the subscriber's device, sound informing about receipt of incoming messages may be turned off, and / or the subscriber may not pay due attention to the received message.

It is also known that to overcome the above disadvantages, additional devices (or their software implementations) can be connected to subscriber devices that process incoming text alerts about emergencies and perform some additional actions.

So, in the USA, an SMS-system for alerting subscribers in case of emergency called “PLAN” (“Personal Localized Alerting Network”) has been developed, capable of working even in a congestion mode of mobile networks. Within the framework of this system, special chips are introduced into subscriber devices [5], which allow them to receive alerts as a priority, even if there are malfunctions on the cellular communication line or many calls are received on the telephone.

The volume of messages in this system does not exceed 90 text characters. Messages are divided into warnings about threats to life and safety, alerts within the search system for abducted or missing children, as well as messages from the President of the United States.

In Israel, in case of danger to the population, a warning system has been created via mobile phones. Subscribers receive alerts regardless of the mobile operator. The device that provides the notification is embedded in the SIM cards of subscriber devices and is activated if necessary [6].

In Russia, a device has been created designed for targeted notification of an emergency (ES) using SMS messages [7]. This device is installed in the subscriber device and provides: reception and reading of the alphanumeric emergency code from incoming SMS; extracting information from data banks stored in the device, depending on the code received; generating an emergency message from the extracted information; launching a vibration signal and an “alarming” melody at the maximum possible volume in the subscriber device; issuance of emergency messages to the built-in tools (screen, speaker).

As the closest analogue (prototype) patented by the author of the invention, the Information Broadcasting System for the US Patent [8] and Russia [9], known by the abbreviation "EDWARDS" (emergency device and warning system for distress and rescue), was selected.

This system actually includes the following essential parts:

- a set of tools for monitoring the current situation (sensors, monitoring services, weather stations, etc.), united under the name of an authoritative source of current control and monitoring;

- a module for receiving information designed to receive information from an authoritative source of current control and observation;

- information transmission module, designed to receive information from the information reception module and generate messages that contain encoded data intended for transmission to subscriber communication devices in at least one pre-selected geographical area;

- subscriber communication devices receiving messages with encoded data and generating notifications for the subscriber corresponding to at least one information component stored in said subscriber communication devices.

Moreover, the subscriber communication device for use in the EDWARDS system contains:

- data storage means in which a plurality of information components are stored, and

- data processing means that are configured to receive messages containing encoded data corresponding to at least one information component stored in the specified communication device, and determining a specific group of subscribers so that when any subscriber communication device receives a message, not intended for a specific group of subscribers, such a message was ignored, otherwise the subscriber communication device generates a notification corresponding to at least one information component.

The method of transmitting information on the selected inventive prototype, includes the steps of:

- obtaining information from an authoritative source of current control and observation;

- the transmission of messages containing encoded data in at least one pre-selected geographical area;

- broadcasting the specified message to communication devices in the specified at least one pre-selected geographical area, while the specified encoded data is configured to match at least one information component stored in these communication devices, and determining a specific group of subscribers so, so that when a subscriber communication device receives a message that is not intended for a particular group of subscribers, such a message will be ignored, otherwise the subscriber’s communication facilities generate a notification corresponding to, but at least one information component.

It is proposed to configure subscriber devices for working with this system using special software products or hardware devices connected to a subscriber device.

However, this invention has significant disadvantages, the elimination of which is directed by the claimed invention. Details of the main deficiencies found are listed below in paragraphs 1-6.

1) The inability of the prototype to work with incompatible subscriber devices and devices that are not subscribers to the system services.

In the description of the invention [8] in paragraph [0117] it is stated that “Communication devices that do not have a code list stored in the subscriber identity module or which are not loaded will not recognize relay messages of mass messages on cells. If the message is accepted, then it will be a meaningless line of code. "

This drawback deprives a large number of subscribers who are not subscribers of the system of the ability to notify using the messages of the prototype system, since even when such a message is received, it will be a “meaningless line of code” for them.

As already mentioned, there are still many “old model” subscriber devices that do not support installing additional programs and connecting third-party devices. Owners of these devices can never become subscribers to the system.

However, even subscribers who have modern devices compatible with the proposed software and hardware can, for one reason or another, unsubscribe from the EDWARDS notification system services. The reasons for the refusal may be technical, economic, psychological, political or any other nature.

2) The need to send different messages to different groups of subscribers in the event of a single emergency.

EDWARDS subscribers are divided into such subscriber groups as the general public, police service, search and rescue service, ambulance, medical and sanitary service, etc., representatives of each of which should receive messages containing correctly formed code data.

Representatives of each group are sent messages with different codes, and if the subscriber communication device receives a message that is not intended for his group, such a message will be ignored.

This drawback makes it necessary to generate, store and send out many different code messages, which can potentially cause unnecessary confusion, which is very undesirable in case of an emergency.

It is also possible to imagine a situation where, as a result of the development of a natural disaster, damage to base stations or other critical transmitting elements of the system is possible, leading to a sudden cessation of messages. Moreover, the subscriber’s device, which still managed to receive a distress alert, even if it wasn’t intended for him, will ignore it and will wait in vain for another message. While a message already received could alert the subscriber and encourage him to perform the necessary actions.

3) The notification system must store, constantly replenish and edit the databases (lists, registers) of subscribers who are representatives of each group of subscribers to the system, as well as who are not subscribers to the system.

At the same time, it is constantly necessary to monitor changes caused by the transition of subscribers from one group to another, the subscription of new subscribers to the services of the system, the suspension or termination of subscription of subscribers.

This creates additional inconvenience and overhead, and can also potentially lead to unnecessary confusion.

4) Problems associated with the overload of wireless cellular networks in an emergency.

A lot of voice calls, usually made by subscribers during an emergency, are so overloading the network that it leads to an unacceptably long delay in transmitting and receiving short messages. At the same time, the main “roll of calls” is observed in the first 3-15 minutes after the start of an emergency and can be so intense (especially in densely populated areas with underdeveloped infrastructure) that it completely disables wireless cellular networks for a long time.

5) Subscribers of the prototype system may incur additional costs for the payment of incoming and outgoing messages (for example, SMS or email) associated with the normal functioning of this system.

The normal functioning of the prototype involves the periodic exchange of information (messages) between the subscriber unit and the warning system. This may cause inconvenience to subscribers related to the payment of incoming and outgoing messages at the current tariffs of their mobile operator. Having experienced such inconvenience and incurring additional costs, subscribers can refuse to use the services of the system, which will lead to a decrease in the number of its subscribers.

At the same time, systems for providing financial transaction services using subscriber devices (mobile phones) are known in the art.

In particular, the system and method according to patent [10] in which a mobile subscriber transmits payment instructions via a mobile phone to a financial operator, who, after authenticating the subscriber, checks the balance on the subscriber’s personal account and, if there are sufficient funds, sends it from the registered short numbers to the subscriber’s mobile phone a set of SMS messages, the total cost of which is not less than the amount received by payment instructions, and the SMS messages received by the subscriber are the basis for the tariff ation in the billing system provider and debits (write-off) of funds to his personal account.

We should also mention the services widely provided by mobile operators that allow you to "easily" transfer money from one mobile number account to any other mobile phone number account. The service is implemented by sending special messages to the address of the mobile operator, containing the number of the recipient, and the transfer amount [11]. Such, for example, are the services “Mobile Transfer” and “Mobile Payment” provided to Beeline subscribers.

6) Notifications issued by subscriber devices do not take into account the possible state of the subscriber and the environment (which can rapidly change) that have developed in the event of an emergency, which can lead to the issuance of erroneous or unrealizable recommendations to the subscriber.

The shortcomings listed above, and other problems, still have not allowed the mass introduction and effective use of alert systems for subscribers of emergencies using short text messages. The author of the claimed invention created it, hoping to eliminate the disadvantages of the prototype and known analogues.

OBJECTS OF THE INVENTION AND SUMMARY OF RESULTS

The general problem to which the invention is directed is to eliminate the identified shortcomings of existing systems for alerting mobile subscribers about emergencies using short text messages.

The technical task is to increase the efficiency of the system and subscriber communication devices, namely:

- providing the ability to alert subscribers who own incompatible subscriber devices and devices that are not subscribers to the system services;

- elimination of the need to send different messages to different groups of subscribers in case of emergency;

- elimination of the need to store databases (lists, registers) of subscribers who are representatives of each group of subscribers to the system, as well as non-subscribers of the system;

- elimination of the need to track changes caused by the transition of subscribers from one group to another, the subscription of new subscribers to the system services, the suspension or termination of subscription of subscribers;

- simplification and automation of the process of generating short text messages containing an alert about a situation;

- ensuring the protection of wireless cellular networks from being overloaded with a multitude of voice calls made by subscribers in case of an emergency (“call shaft” in the first 3-15 minutes after the onset of the situation);

- replenishment of expenses of subscribers of the system for the payment of incoming and outgoing messages associated with the normal functioning of the system;

- ensuring that subscriber devices take into account the state of the subscriber and the environment that developed during an emergency when they issue notifications and recommendations.

SUMMARY OF THE INVENTION

The overall objective is achieved by solving subtasks corresponding to various aspects of this invention.

The aspects of the claimed invention described below should be considered in conjunction aimed at solving a single technical problem and the embodiment of the inventive concept of the author. The joint implementation of aspects of the invention allows to achieve the desired results.

As an explanation, we mention that the paragraph numbers of the appended claims generally correspond to the numbers of aspects of the invention listed below.

According to the first aspect of the invention, the warning system owns a single set of telephone numbers registered and maintained by all local, regional and global mobile operators operating in the territory of the deployment of the system.

At the same time, the information receiving module is configured to bring the information received from the monitoring tools for the current situation to a single format and dimension.

The information transfer module is configured to transmit absolutely identical copies of the message generated by it to all subscriber communication devices located in the geographical area of emergency deployment, regardless of which group the subscribers belong to and whether they are subscribers to the system or not;

The information transfer module also contains a data bank of text segments in a limited controlled version of the natural language, mainly used by residents of the geographical area to be notified, each stored text segment carries information about a specific parameter of a possible emergency.

And also, the information transfer module contains a trained multilayer artificial neural network (ANN), a plurality of information inputs, which corresponds to a plurality of information channels for monitoring the current situation, and a plurality of outputs corresponds to a plurality of text segments stored in a text segment data bank, used to generate short text messages intended for transmission to subscriber communication devices.

And also the transmission module contains a block for the final generation of a text message located between the outputs of the ANN and a data segment data bank, configured to generate a message intended for transmission to subscriber communication devices by:

- extracting from the data bank those text segments that correspond to the activated ANN outputs;

- separation of text segments using the same grammar separator characters;

- restrictions on the length of the text of the generated message so that it does not exceed the maximum allowable length for one short text message.

The information transfer module contains means for coordinating the beginning of the notification process with the authorized government body and indicating to it the geographical area of the probable emergency situation to be notified, connected to the final text message generation unit.

And also the information transfer module contains built-in training tools for the ANN and its configuration, designed to organize the process of training the ANN based on the opinion of a group of competent experts, including:

- A control processor with a set of commands and customizable software;

- random number generator;

- generator of visual and audio images;

- a database of images and sounds;

- memory block;

- means for displaying and entering information with the ability to read expert opinions and store weighting factors assigned to experts;

- a block for calculating the majority of votes of the interviewed experts, taking into account the assigned weighting factors;

- block correction of synaptic weights of ANN neurons in accordance with a given correction step;

compatible interfaces for connecting to the ANN.

At the same time, the functional connections between the above means of training and configuring the ANN are organized so that the information inputs / outputs of the control processor are connected to the corresponding outputs / inputs of the random number generator, the generator of visual and audio images, the memory unit, the correction unit of the synaptic weights of the ANN neurons, the majority counting unit votes of the interviewed experts, a databank of text segments.

In this case, the information inputs / outputs of the majority vote block of the interviewed experts are connected to the outputs / inputs of the information display and input devices.

Also, the information output / input of the image and sound data bank is connected to the input / output of the generator of visual and audio images.

The information output of the memory unit is connected to the input of the correction unit of the synaptic weights of the ANN neurons.

The indicated unit for the correction of synaptic weights of ANN neurons is also connected to the following interfaces for connecting to a trained neural network:

- an output for supplying training input signals to the neural network;

- output for correction of the vector of synaptic weights of neurons of the trained neural network;

- an input for reading the output signals of a trained neural network.

According to a second aspect of the invention, a broadcast notification system is characterized by the following features.

Subscriber communication devices belonging to subscribers of the system further comprise a unit for tracking changes in the balance on the account of the subscriber’s mobile number caused by the exchange of messages between the subscriber device and the notification system.

Information inputs / outputs of the indicated block are connected to the system of receiving / sending messages of the subscriber device.

The specified unit is configured to periodically automatically request information from the subscriber’s mobile operator about the account status of the subscriber’s mobile number and, having registered debit funds from the account caused by the exchange of messages between the subscriber unit and the notification system, generates a message with a unique confirmation code.

This message contains at least the following information:

- the amount of money to replenish the subscriber's account;

- number identifying the subscriber device;

- date and time of message generation.

Then the unit prompts the device to send this message to a specific short number belonging to the notification system.

Moreover, the system contains a module that automatically performs the following actions:

- receipt of messages from subscriber communication devices containing unique confirmation codes to a specific short telephone number belonging to the system;

- verification and analysis of these confirmation codes;

- replenishment of the accounts of mobile numbers of subscribers for an amount equal to or slightly higher than debited from their accounts due to the exchange of information between subscriber devices and the notification system, by transferring funds from the account of the telephone number belonging to the notification system to the account of the mobile number of the subscriber.

According to a third aspect of the invention, a subscriber communication device belonging to a subscriber of a patented system comprises the following elements.

The data processing unit, configured to receive short text messages, extracting text segments from a message received from telephone numbers of the warning system, in a limited controlled version of the natural language, each of which carries information about a specific emergency parameter, and is indicated by the same grammatical characters- delimiters, and converting each text segment to binary.

A multilayer artificial neural network (ANN), the set of information inputs of which corresponds to the set of text segments contained in short text messages coming from the telephone numbers of the warning system, and the set of outputs corresponds to the set of information components contained in the data storage unit.

The unit for collecting individual information about the subscriber, configured to display questions on the built-in screen of the subscriber device and read user responses.

A unit for collecting current information about the status of the subscriber and the environment using the built-in sensors and systems of the subscriber device.

The ANN training and configuration unit configured to record and store the collected individual and current information inside the ANN by changing the ANN settings corresponding to the number of ANN layers, the structure of neuron connections and the values of synaptic weights so that the ANN contained in the subscriber communication device upon receipt of a message from the warning system, the outputs corresponding to the set of information components most suitable for a particular subscriber and the environment are activated, In spite of the fact that the content of messages sent by the system alert to all subscribers in the event of an emergency, the same.

A notification program generation unit located between the ANN outputs and the data storage unit, configured to generate a notification to the subscriber from a plurality of information components corresponding to a plurality of activated ANN information outputs.

A block that prompts the subscriber device to perform certain actions corresponding to the generated subscriber alert program.

In this case, the functional connections between the elements of the subscriber communication device are organized so that the input of the data processing unit is connected to the system for receiving / sending messages of the subscriber device, and the outputs are connected to the information inputs of the ANN.

The ANS training and configuration unit is connected to the following interfaces connected to the ANN:

- an output for supplying input signals to the ANN;

- an exit for change of settings of ANN;

- an input for reading the output signals of the ANN.

The ANN training and configuration unit is also connected to the unit for collecting individual information about the subscriber and the unit for collecting current information about the status of the subscriber and the environment.

The unit for collecting individual information about the subscriber is also connected to the built-in screen of the subscriber device and the keyboard.

The unit for collecting current information about the status of the subscriber and the environment is connected to the built-in sensors and systems of the subscriber device.

The notification program generation unit is connected to the outputs of the ANN and the data storage unit, the output of the notification program generation unit is connected to the input of the unit that prompts the subscriber unit to perform certain actions.

The unit that prompts the subscriber device to perform certain actions is connected to the following elements built into the subscriber device: vibration alarm, speaker, LED indicators and flash of the built-in camera, a screen with a keyboard.

According to a fourth aspect of the invention, a subscriber subscriber communication device of a system, inter alia, comprises the following means of preventing overload of a wireless communication network.

A random number generator connected to a block that prompts the device to perform certain actions.

The unit that prompts the device to perform certain actions is configured to block, in the subscriber communication device, the functions of making outgoing voice calls and sending messages for a period of 3 to 15 minutes from the moment of receiving the corresponding text message, which is randomly selected when receiving a short text message from the notification system in the specified range using a random number generator.

And also this unit is configured to display a notification on the screen and speaker of the subscriber device informing about the impossibility of outgoing voice calls and sending messages, as well as about the time remaining until the end of the blocking.

The following aspects of the invention will relate to methods of functioning of the claimed system and devices included in its composition.

According to a fifth aspect of the invention, a method for operating the system in the mode of generating and transmitting short text messages to subscriber communication devices is patented. This method includes the following steps:

- register and maintain a single set of telephone numbers belonging to the warning system, with all local, regional and global mobile operators operating in the territory of the deployment of the system;

- receive and lead to a single format and dimension information from means of tracking the current situation;

- provide information from the means of monitoring the current situation to the information inputs of a pre-trained multilayer artificial neural network (ANN), many of which correspond to many information channels of means for monitoring the current situation;

- generate, using the ANN, a single short text message in a limited controlled version of the natural language, mainly used by residents of the geographic area to be notified, from text segments, many of which are stored in the text segment data bank, and correspond to many ANN outputs;

- the message text is divided into a certain number of segments using the same grammar separator characters, and each segment carries information about a specific emergency setting;

- the total length of the text is chosen so that it does not exceed the maximum allowable length for one short text message;

- determine the geographical area of the likely deployment of an emergency;

- coordinate the beginning of the notification process with the authorized authority;

- transmit the same copies of a short text message previously generated using the ANN to all subscriber communication devices located in a pre-selected geographical area of emergency deployment, regardless of which group the subscribers belong to and whether they are subscribers to the system or not;

- receive the indicated messages to subscriber communication devices;

- receive messages from subscriber communication devices containing unique confirmation codes to a specific short telephone number;

- read and verify the specified confirmation codes;

- replenish the accounts of mobile numbers of subscribers for an amount equal to or slightly higher than debited from their accounts due to the exchange of information between subscriber devices and the notification system, by transferring funds from the account of the telephone number belonging to the notification system to the accounts of mobile numbers of subscribers.

According to a sixth aspect of the invention, a method for the operation of subscriber communication devices belonging to subscribers of the system when a message about an emergency occurs from the notification system is patented. This method includes the following steps:

- receive a short text message from the warning system;

- extract text segments from the message, each of which carries information about a specific emergency parameter, and is indicated using the same grammatical separator characters;

- translate each text segment into binary;

- serve the selected text segments to the information inputs of a multilayer artificial neural network (ANN) corresponding to the text segments of the message;

- read signals from the multiple outputs of the ANN corresponding to the set of information components;

- extract those information components from the data storage means that correspond to the activated information outputs of the ANN;

- start the execution of certain actions of the subscriber device to alert the subscriber corresponding to the extracted information components and the activated information outputs of the ANN;

- randomly select a period of time from a range of 3-15 minutes using a random number generator;

- block in the subscriber communication device the functions of making outgoing voice calls and sending messages for a period of time corresponding to the period selected in the previous step;

- display on the screen and speaker of the subscriber device information about the impossibility of outgoing voice calls and sending messages, as well as about the time remaining until the end of the block;

- restore the normal operation of the subscriber device in everyday use;

- track changes in the balance on the account of the mobile number of the subscriber caused by the exchange of messages between the subscriber device and the warning system;

- replenish the account balance by an amount equal to or slightly higher than the written-off, by generating and sending to a specific short number a message with a unique confirmation code containing at least the following information:

- the amount of money to replenish the subscriber’s account,

- number identifying the subscriber device,

- date and time of message generation.

According to a seventh aspect of the invention, the method for operating subscriber units described in the sixth aspect of the invention can be improved by the fact that it further includes the following steps:

- register a clear difference between the actions actually performed by the subscriber and the recommendations issued by the communication device when a message is received from the notification system;

- display a request to the subscriber whether he considers the actions actually performed by him to be correct;

- fix the actions actually performed by the subscriber;

- make the assessment by the subscriber of the correctness of the actually performed actions;

- they train and configure the ANN stored in the subscriber communication device based on the actions actually recorded by the subscriber taking into account the assessment of their correctness in order to eliminate the differences between the recommendations issued and the actions actually performed by the subscriber in the event of identical emergency situations in the future.

The invention will now be described with reference to the accompanying drawings. The following description is intended to disclose preferred embodiments of the invention.

In the following description, the means and methods known at the filing date of the application will be described in detail, allowing to realize the above characteristics of the main aspects of the invention and the attached claims.

BRIEF DESCRIPTION OF THE DRAWINGS

The description is illustrated by figures of graphic materials illustrating a preferred embodiment of the invention.

The figure 1 schematically shows the settlement 1 (a city with a population of approximately 30,000 inhabitants), located in a mountainous area. This settlement is located in a valley bounded on two sides by mountain slopes, dangerous from the point of view of the sudden occurrence and flow of mudflows. In figure 1, the western and eastern slopes of the mountains are indicated, respectively, under the numbers 2 and 3.

The figure 2 shows the General deployment scheme of one of the preferred embodiments of the patented invention in the above-described city 1 in order to alert residents about emergency situations associated with the occurrence and debris flow.

In figure 2, the following notation is used: 4, 5, 6, 7 — a set of means for monitoring the current situation on the western slope; 8, 9, 10, 11 - a set of tracking tools for the current situation on the eastern slope; 12, 13 - information channels for monitoring the current situation, respectively, on the western and eastern slopes; 14 - a module for receiving information from monitoring tools for the current situation; 15 - information channel for data transmission, implemented on the basis of fiber optic cable; 16 - information transfer module; 17 - information channel for transmitting data to a GSM network; 18 - GSM-network mobile wireless communications deployed in the described locality; 19 - subscriber units of city residents; 20 - actually Broadcast system for alerting mobile subscribers about emergencies.

Figure 3 shows in more detail the main essential parts of the information transfer module 16. Figure 3 uses the following notation: 21 — artificial neural network (ANN) of the “multilayer perceptron” type; 22 - block the final generation of a text message; 23 - databank of text segments; 24 - memory block; 25 is a block correction of the synaptic weights of ANN neurons in accordance with a given correction step; 26 - automated workstation (AWS) of the operational on-duty city center for crisis management of the Ministry of Emergency Situations (MCC CCM); 27 - random number generator; 28 - control processor with a set of commands and custom software; 29 - generator of visual and audio images; 30 - a database of images and sounds; 31 - majority block counting the majority of votes of interviewed experts; 32, 33, 34 - AWP of experts, including means for displaying and entering information, means for reading expert opinions and means for storing weighting factors assigned to experts; 35, 36, 37 - compatible interfaces for connecting to the ANN.

The figure 4 shows a diagram of the main blocks of the subscriber communication device 19, owned by subscribers of the warning system 20, which shows the following essential parts: 38 - data processing unit, configured to receive short text messages from the numbers of the warning system; 39 - artificial neural network (ANN) type "multilayer perceptron"; 40 - block training and configuration of the ANN; 41 - unit for collecting current information about the status of the subscriber and the environment; 42, 43, 44 - built-in sensors and subscriber device systems; 45 - block tracking changes in the balance on the account of the mobile number of the subscriber; 46 - block generating a subscriber alert program; 47 — a data storage unit in which a plurality of information components are stored; 48 is a block that prompts the subscriber device to perform certain actions; 49 - random number generator; 50 - unit for collecting individual information about the subscriber; 51 - system for receiving / sending SMS messages; 52 - a memory block for storing a queue of incoming SMS messages; 53 - a memory unit for storing a queue of outgoing SMS messages; 54 - vibration alarm; 55 - speaker; 56 - LED indicators and flash integrated camera; 57 - touch screen with on-screen keyboard.

Figure 5 shows in more detail the main elements of the GSM network of mobile wireless communications 18. Figure 5 uses the following English-language abbreviations: MSC - Mobile Network Switching Center; BSC - Base Station Controller; CBS - Cell Broadcast Center BTS - the base station of the network (Base Transceiver Station).

The figure 6 shows a block diagram of the operation of the module 14 to bring the information received from the means of tracking the current situation to a single dimension. As an example, information obtained from temperature sensors was used.

The figure 7 shows a block diagram of the operation of the unit for the final generation of a text message 22 to form a message line from text segments.

The figure 8 shows a block diagram of the workstation AWP 26 on-duty duty MSC EMERCOM (decision maker, decision-maker) when activating the outputs of ANN 21 and block 22.

Figure 9 illustrates the operation of the unit for tracking changes in the balance on the account of the mobile number of the subscriber 45 in the subscriber communication device.

Figure 10 schematically illustrates the operation of means for receiving messages from subscriber devices containing unique confirmation codes and replenishment of accounts of mobile numbers of subscribers.

The figure 11 shows a block diagram of the unit for tracking changes in the balance on the account of the mobile number of the subscriber 45.

Figure 12 shows a block diagram of the operation of a module that receives messages from subscriber devices containing unique confirmation codes.

The block diagram of the algorithm of operation of the data processing unit 38 of the subscriber device configured to receive short text messages from the notification system numbers is shown in FIG. 13.

Figure 14 illustrates the numbering of the inputs of the ANN 39 of the subscriber unit, many of which correspond to many text segments (converted to binary numbers), and the many outputs of the ANN are numbered so that the numbers of the outputs correspond to the numbers of information components stored in the data bank 47.

The figure 15 shows a block diagram of the operation of the unit 48, prompting the subscriber device to perform certain actions, in order to prevent overloading the communication network by blocking the communication device 19 functions of making outgoing calls and sending messages.

In order not to overload the graphic materials with redundant notations and to maintain their visibility, some parts of the patented system and its constituent devices, as well as the connections between them, were not shown on the presented figures of graphic materials. If necessary, these parts and the relationships between them can be restored by specialists on the basis of a textual description of the invention, prior art and common sense.

Many blocks that perform similar functions, the number of which may vary depending on the particular implementation of the invention, are shown in the figures of the drawings using the ellipsis symbol "...". From the following textual description it will be clear what this means in each case.

DETAILED DESCRIPTION OF THE INVENTION

General Provisions

The author implies that the use of the proposed system and the devices included in its composition must comply with the legislation, standards and customs prevailing in the territory of the deployment of the system.

In city 1, a GSM mobile wireless communications network 18 has been deployed and is constantly operating, including BTS base stations that provide mobile communications throughout the village. The city has administrative and residential buildings of various storeys, facilities for cultural events, service points, trade, etc.

The distribution of mobile devices (of all types) reaches 100%, i.e. almost every resident of the city has a mobile phone. At the same time, about 40% of residents remain using old-style subscriber devices that do not support the installation of additional programs and the connection of third-party devices.

The remaining residents (60%) have devices that can technically become subscribers to the warning system. Of these, about half are already subscribers to the system, the rest - not yet. Thus, the number of permanent subscribers to the system is approximately 30% of the city’s population, i.e. about 9000 inhabitants.

In the figures of graphic materials, number 19 denotes all subscriber devices of city residents, including “old-style” devices, devices of system subscribers and not being such. In the future, it will be clear from the description which device we are talking about.

The author considers several possible options for subscribing subscribers to system services. It could be:

- downloading special software developed for the popular operating systems of smartphones “Google Android”, “Apple iOS”, “Microsoft Windows Phone”, and placed in the Internet services of the mobile applications “Google Play”, “Arp Store”, “Windows Phone Store” ";

- free replacement of SIM cards of the GSM standard (Subscriber Identification Module) with SIM cards that are in the form of ordinary SIM cards, but containing microcircuits, memory modules and special microprograms that provide the necessary functionality, in accordance with a technology similar to the SIM Tool Kit "(STK);

- connection to the subscriber device via external mini-USB (or micro-USB, or through a slot for Digital Memory Card) external devices that provide the necessary functionality;

- integration into subscriber communication devices of various microcircuits and / or software in specialized workshops or at the factory.

The possibility of sharing several of the above options is also being considered. Thus, the logic of the work of the blocks of the subscriber device (shown in Fig. 4), embedded in them, can be implemented using both hardware and software.

The specific type of logic used in a particular block of the subscriber device 19 does not fundamentally affect the operation of the device. The achievement of the technical result also does not depend on the type of implementation of the logic of the individual blocks.

Note that the warning system does not monitor or monitor whether a particular device is a subscriber of the system or not. The list (database, register) of all system subscribers is not compiled and not maintained. In the event of an emergency, the same messages are sent to all subscriber devices (including “old-style” devices and devices that are not subscribers to the system) located in the emergency deployment area.

In fact, for a user, a subscription consists only in voluntarily changing the functionality of his own subscriber device, in accordance with known rules, which he can provide using the subscription options for system services described above. At the same time, he may unsubscribe by removing the corresponding software and / or hardware from his communication device.

A possible update of the subscriber device provision is mainly carried out by receiving instructions for setting up the ANN 39 contained in the subscriber device. Such instructions may include the recommended number of ANN layers, neural neural connections, and synaptic balance values. Instructions can be automatically received from the update server using incoming SMS messages or when a subscriber device is connected to the mobile Internet, for example, via Wi-Fi.

As an explanation, we mention that the ANN is a hardware and / or software system built on the principle of the organization and functioning of nerve cell networks of living organisms. It is a system of interacting artificial neurons (simple computing elements) connected to a sufficiently large network with several layers. By determining the proper weights of interneuronal (synaptic) connections (“training”), ANNs can achieve effective solutions to various problems. Learning ability is one of the benefits of neural networks.

Necessary explanations to the first aspect of the invention.

Local representatives of the Ministry of Emergencies register and maintain a single set of telephone numbers belonging to the notification system 20 at all local, regional and global mobile operators operating in the city 1, where the system 20 is deployed.

This set of numbers includes at least one number for sending short text messages with an alert about an emergency, and another short number for receiving messages sent by subscriber devices in order to recover money on the accounts of mobile numbers of subscribers debited from for the exchange of messages between the subscriber device 19 and the warning system 20.

According to FIG. 1 and FIG. 2 on the western slope of mountain 2 adjacent to city 1, there is a set of monitoring tools for the current situation, including a group of seismic activity sensors 4, soil movement sensors 5, soil moisture sensors 6, temperature sensors 7. This set of tools can be expanded, by adding new groups of sensors, for example, means for measuring wind speed or noise level, as indicated in FIG. 2 using the ellipsis symbol "...".

On the eastern slope 3 there is an identical set of monitoring tools for the current situation, including a group of seismic activity sensors 8, soil motion sensors 9, soil moisture sensors 10, temperature sensors 11. This set of tools can also be expanded by adding new sensor groups, which indicated in FIG. 2 using the ellipsis symbol "...".

Means for monitoring the current situation regularly transmit the readings of the respective sensors to the information receiving module 14 using information channels 12 and 13, implemented mainly on the basis of wireless radio channels.

Module 14 should bring the received information to a single format and dimension, because The sensors used have individual characteristics, caused, in particular, by the fact that they can be produced by different manufacturers and at different times, use specific units of measurement, and have different thresholds for false response and errors. Another reason is that the information supplied by it to the inputs of ANN 21 must be strictly standardized and normalized.

Therefore, module 14 is configured to bring the information received from the means of tracking the current situation to a single format and dimension. To implement this function, the module is supposed to be executed in the form of a programmable multifunctional tool (a computer that implements a certain algorithm of actions).

For example, for the information received from temperature sensors 7 and 11, this function of module 14 is as follows (a block diagram of the algorithm is shown in Fig. 6).

First, bring the numerical temperature values obtained from different sensors working on the Celsius and Fahrenheit scales to a single measurement system on the Celsius scale.

For this, the readings of sensors measuring temperature in Fahrenheit (t _F ) are converted to degrees Celsius (t _C ) according to the well-known formula:

And the readings of sensors measuring temperature in Celsius remain unchanged or are multiplied by unity (which, in essence, is the same).

Secondly, the normalization of the information supplied to the inputs of the ANN 21, requires that the numerical values of the supplied parameters are in the range from 0 to ± 1. To do this, all temperature indicators should be divided by 100. Thus, if initially we had sensor readings equal to 35 and 43 degrees Celsius, then after dividing by 100 we will get readings equal to 0.35 and 0.43, respectively.

Conversion of information received from the remaining sensors (4-6 and 8-10) to a single format and dimension occurs using algorithms similar to the order described above for temperature sensors.

As a result, module 14 provides information in the following form:

X = <fw ₁ , fw ₂ , ..., fw _n ; qw ₁ , qw ₂ , ..., qw _n ; hw ₁ , hw ₂ , ..., hw _n ; vw ₁ , vw ₂ , ..., vw _n ; fe ₁ , fe ₂ , ..., fe _n ; qe ₁ , qe ₂ , ..., qe _n ; he ₁ , he ₂ , ..., he _n ; ve ₁ , ve ₂ , ..., ve _n >,

where fw ₁ , fw ₂ , ..., fw _n is the set of readings of sensors 4;

where qw ₁ , qw ₂ , ..., qw _n is the set of readings of sensors 5;

where hw ₁ , hw ₂ , ..., hw _n is the set of readings of sensors 6;

where vw ₁ , vw ₂ , ..., vw _n is the set of readings of sensors 7;

where fe ₁ , fe ₂ , ..., fe _n - the set of readings of sensors 8;

where qe ₁ , qe ₂ , ..., qe _n is the set of readings of sensors 9;

where he ₁ , he ₂ , ..., he _n is the set of readings of sensors 10;

where ve ₁ , ve ₂ , ..., ve _n is the set of readings of the sensors 11.

Where n is the number of system sensors included in the corresponding sensor group.

And then it transfers it to the inputs of the ANN 21 located in the information transfer module 16, using the information channel based on the fiber optic cable 15.

Modules 14 and 16, as a whole, are separate seismic-resistant buildings in the administrative (central) part of the city, equipped with appropriate equipment and maintenance personnel.

The information transfer module 16 is configured to transmit absolutely identical copies of the message generated by it to all subscriber communication devices located in the geographical area of emergency deployment, regardless of which group the subscribers belong to and whether they are subscribers to the system or not.

This function of module 16 is realized by the fact that block 22 interacts with AWP 26, which ensures that the person in charge of the geographical area indicates the probable emergency situation to be notified.

And also by the fact that block 22 transmits the generated message indicating the geographical area of the emergency deployment to be notified to the GSM mobile wireless network 18 containing the Cell Broadcast element: SHS - broadcast center.

SHS stores information about the location of existing base stations and their geographical location. SHS is connected to base station controllers BSC (Base Station Controller).

SHS stores information on the correspondence of the geographical area specified by block 22 and AWP 26 with the location of existing base stations and sends a message only to the necessary base station controllers (BSC), which, in turn, already redirect this message to the necessary base stations to transmit identical copies of the message to all active subscriber communication devices located in the geographical area of emergency deployment.

By “limited controlled version of a natural language, primarily used by residents of a geographic area to be notified”, is meant a subset of the natural language obtained by limiting its vocabulary, grammar and speech speed in order to reduce its ambiguity and complexity.

The author suggests that in English-speaking countries, in the form of a controlled language (“CNL”), variants of the so-called “Special English” can be used, which is a controlled version of the English language specially designed for broadcasting stations “Voice of America”. Special English and its variations, such as Basic English and Simple English, are characterized by the use of simple short sentences and a limited vocabulary.

The simplest example of a databank of text segments 23, implemented in the form of a table storing the correspondence of each output of the ANN 21 and the text segment, is shown below in table 1.

The author also means the possibility of storing text segments in transliteration (a way of writing Russian text in Latin letters, as well as numbers and other available characters). Using transliteration can increase the total length of the generated message to 160 characters (restriction for SMS messages - no more than 160 characters of the Latin alphabet). While for the Russian language, the message length cannot exceed 70 characters.

It should be clarified that the artificial neural network ANN 21 is implemented as an ANN of the “multilayer perceptron” type with the number of hidden layers of at least two, which can be implemented both hardware and software (for example, as a dll module of a connected library).

The author also considers various options for implementing system 20, in which, as an ANN (in blocks 21 and 39), not only neural networks of the “multilayer perceptron” type are used, but also their varieties such as “Kohonen networks”, “Hopfield networks” etc.

The set of information inputs of ANN 21 corresponds to the set of information channels for monitoring the current situation, converted by module 14 to a single format and dimension.

Many outputs of the ANN are numbered so that the numbers of the outputs correspond to the numbers of the text segments stored in the database 23.

The final text message generation unit 22, located between the ANN outputs and the text segment data bank, is a multifunctional tool that implements certain algorithms. The flowchart is shown in FIG. 7.

Block 22 is configured to extract from the data bank 23 those text segments whose numbers correspond to the numbers of the activated outputs of ANN 21. It is assumed that block 22 extracts text segments from 23 in the same sequence in which the corresponding outputs of ANN 21 are activated.

Next, block 22 generates a text string, arranging the extracted text segments in the order in which they are extracted, and inserts the same grammar separator between all the extracted text segments. In a preferred embodiment of the invention, a mark is used for this; "semicolon".

Then, block 22 estimates the length of the generated text string. A function similar to the standard Length () function of the Pascal programming language can be used for this. If it exceeds the maximum permissible length for one short text message (70 characters in Russian or 160 characters in Latin), then the last text segment and the separator character are discarded.

Once again, block 22 estimates the length of the generated text string. If it still exceeds the permissible length, then the process repeats. Otherwise, the message text is considered formed.

In the context of the claimed invention, the joint use of properly trained ANN and a databank of text segments can significantly facilitate and automate the process of generating short text messages in a limited version of a natural language, containing an alert about a situation.

Means ensuring coordination of the beginning of the notification process with the authorized government body and indicating to them the geographical area of the probable emergency situation to be notified are implemented in the form of a 26 - automated workstation (AWP) of the operational on-duty city crisis management center of the Ministry of Emergency Situations (CCC MES) )

It is assumed that the person on duty is a decision maker (decision maker) endowed with the necessary powers.

AWP 26 is a software and hardware tool that provides interaction with a person by entering information (through the keyboard and computer mouse), its output to the monitor screen, sound speakers, as well as storage and processing of information.

AWP 26 provides a conclusion to the person in charge of the fact of an emergency (determined by activating the outputs of the ANN 21 and block 22) and the text message generated by block 22, which is proposed for transmission to subscriber devices 19.

If necessary, the operational duty officer from his workstation can make adjustments to the generated text message, determine the area of probable deployment of an emergency, and confirm his consent to start the process of notifying city residents 1.

AWP 26 stores in the form of a file information about the location of existing base stations and their geographical location, identical to that stored in the CBC - Cell Broadcast broadcast center.

The simplest version of such a file is shown in table 2 below.

AWP 26 provides access for the responsible person to this information, and makes it possible for him to indicate the geographical area of the probable deployment of an emergency to be notified, as well as establishing correspondence between the notification area and the location of base stations.

Having received consent from AWP 26 to start the notification process, block 22 is configured to transmit the generated text message via information channel 17 to the GSM network 18 deployed in city 1, indicating the base stations to which the message should be transmitted, in a form compatible with the center format Broadcast Cell Broadcast.

The flowchart of the above steps is shown schematically in FIG. 8.

At the same time, the author believes that, in the absence of any actions on the part of the decision-maker to confirm consent to the start of the notification process within a reasonable period of time, block 22 can independently transmit the generated text message to the GSM network indicating the entire city as notification geographical area 1 , without receiving confirmation from the decision maker (the principle "silence of the decision maker is a sign of consent").

Next, we explain the possible options for implementing the ANS training tools and its configuration.

A control processor with a set of commands and customizable software can be implemented as an ARM architecture processor running on an operating system such as Linux.

The random number generator is implemented using thermal noise as the source of entropy due to the thermal motion of charge carriers in the conductor, which induces a fluctuating potential difference at the ends of the conductor.

The generator of visual and audio images 29, based on the information components stored in the data bank 30, is configured to create visual and audio images describing possible gathering of mudflows defined by the generated signals corresponding to the readings of sensors 4-11, with the greatest possible visualization and reliability.

In fact, the information components stored in the image and sound data bank 30 are computer files that store information about three-dimensional images (stereo images) of the descent of all kinds of mudflows and the consequences of their impact on various objects, in the form of short-term sequences of interchangeable frames (video clips) accompanied by characteristic sounds.

The memory unit 24 may be a drive on hard magnetic disks ("Winchester") or on solid-state memory elements ("solid-state drive", SSD).

By “means of displaying and inputting information with the ability to read expert opinions and store weight coefficients assigned to experts” we mean experts' workstations (32-34) implemented on the basis of a personal computer equipped with a 3-D stereo display, sound-emitting speakers, a keyboard, and a mouse-type manipulator .

The author also suggests that the above elements of expert workstations can be implemented using more modern technologies. Such as trackball; touchpad projection keyboard; aerosol display ("free-space display"), projecting visual 3D images in the amount of free space filled with the aerosol mixture; bi-directional neurocomputer interface for direct exchange of information with the brain of an expert, etc.

Each expert is uniquely identified by the workstation (implemented on the basis of a personal computer) when a unique “login-password” pair is introduced.

At the first use session, each expert enters data about himself into the system, answering computer test questions, on the basis of which each of them is assigned weight coefficients that correspond to their experience and qualifications. For example, if the answer to the question “Are you a candidate of science?” Is positive, then the “weight” of the expert increases by 0.2 units, if the answer is no, then the “weight” remains unchanged, etc.

The methods for assigning different weighting factors to experts depending on their experience and qualifications in order to ensure that the votes of different experts have different “weights” are well known. They can be implemented on the basis of computer routines and do not require additional description.

The coefficients are stored in the memory of the workstation and during repeated sessions they are assigned to the expert automatically upon identification, already without going through the polling procedure.

The majority vote block of the interviewed experts, taking into account the assigned weighting factors 31, represents a programmable multifunctional majority element calculating the result of the experts' vote.

It summarizes the votes of all experts, previously multiplied by the weighting factors corresponding to the experts. If the total number of votes cast “for” is greater than the similar number of votes “against”, then the decision shall be considered adopted, otherwise, rejected.

The correction unit for the synaptic weights of the ANN 25 neurons in accordance with a given correction step is configured to correct, as the ANN 21 is trained, the perceptron of the weight of the synaptic connections w (n) ANN in accordance with the following algorithm:

The difference Y (n) -Y ′ (n) in formula (2) is the error ε (n) between the reference (given) signal Y (n) and the actual output signal Y ′ (n), and the parameter η for 0 <η < 1 is a correction step. Moreover, X (n) = [x ₁ , ..., x _N ] ^T represents the vector of input ANN signals.

It is proved that algorithm (2) converges with iteration n ₀ , i.e.

According to the convergence theorem of the learning algorithm of the perceptron with a fixed increment, this indicates the end of training.

Compatible interfaces for connecting to ANNs include:

- output 35 for supplying training input signals to the ANN;

- output 36 for correction of the synaptic weights vector of ANN neurons;

- input 37 for reading the output signals of the trained ANN.

The warning system 20 can operate in three main modes:

- learning mode multilayer ANN;

- daily functioning mode;

- The mode of generation and transmission of short text messages to subscriber communication devices.

We describe the operation of the system in each mode in more detail.

The warning system 20, in the training mode of the multilayer ANN 21 contained in the information transfer module 16, operates as follows.

First, they form a group of competent experts, which may consist of specialists of various profiles (psychologists, linguists, employees of rescue services, weather forecasters, scientists in the field of geological and mineralogical sciences, etc.). Then, these experts are placed behind workstations 32, 33, 34, the number of which corresponds to the number of experts, and can be increased if necessary by adding new workstations, which is indicated in FIG. 3 using the ellipsis characters "...".

During the first use session, each expert enters data about himself into the system, answering the test questions, on the basis of which each of them is assigned weight coefficients corresponding to their experience and qualifications. The coefficients are remembered and during repeated sessions they are assigned to the expert automatically when he is identified by the login-password pair without going through the polling procedure.

Experts, on the basis of one of the ANN training programs contained in the control processor 28, determine the number of training signals necessary for training corresponding to the signals from the tracking means for the current situation. The number of training signals can be calculated by the formula:

where w is the number of adjustable parameters of the ANN;

e - permissible classification error.

Based on their knowledge, experts indicate the acceptable range of variation of training signals, i.e. in fact, indicate in what range the readings of sensors 4-11 mounted on slopes 2 and 3 can vary.

The decisions of the expert group at all stages and on all issues are selected by counting the majority of votes, taking into account the weighting factors assigned to each expert, in the majority block of vote counting of the interviewed experts 31.

Next, the random number generator 27, at the direction of the control processor 28, generates a previously determined number of training signals corresponding to the readings of sensors 4-11, belonging to an acceptable range of variation of the training signals.

The generator of visual and audio images 29, based on the information components stored in the databank 30, creates visual and audio images describing possible debris flows, defined by the generated signals, corresponding to the readings of sensors 4-11, with the greatest possible visualization and reliability.

Block 28 demonstrates to the experts on their AWP 32-34 generated signals that correspond to the readings of sensors 4-11, and images that clearly describe the development and possible consequences of the mudflows corresponding to the generated signals.

Each expert, based on his knowledge, for each generated situation (mudflow), selects from the bank of text segments 23 that set of text segments, which, in his opinion, is most suitable for generating a short text message intended for further transmission to subscriber communication devices 19 city dwellers 1, when a mudflow was shown to him.

Block 31, by counting votes, taking into account the weights assigned to each expert, determines the set of text segments that satisfies the opinion of the majority of interviewed experts.

As a result, the control processor 28 stores in the memory block 24 the generated training signals and numbers corresponding to text segments, which, according to experts, are most suitable for generating a short text message, in the form of pairs, for example, of this kind:

Z = [X, Y],

where X is the generated training signals that correspond to the readings of sensors 4-11 reduced to a single dimension and format, belonging to a certain allowable range;

Y = <m ₁ , m ₂ , ...> are the numbers of the corresponding text segments, which, according to experts, are most suitable for generating a short text message.

After the formation of the number of training pairs Z, equal to or greater than the number S from formula (4), necessary for training the ANN, the control processor 28 sends a signal to the synaptic weights correction unit of neurons 25 to start the process of direct training of the ANN 21.

Upon receiving the signal, the synaptic balance correction unit 25 reads the recorded Z pairs from the memory unit 24, supplies the training signals X to the information inputs of the trained ANN via interface 35 and reads the output signals of the trained ANN Y ′ through interface 37.

Comparing the output signals of the trained ANN Y ′ with the corresponding numbers of text segments Y obtained from the memory unit 24, the synaptic balance correction unit 25 calculates the error between the Y signal and the actual output signal Y ′ and corrects the synaptic weights of the ANN 21 neurons via the interface 36 in accordance with the specified correction step (see formula 2) until completion of training (see formula 3).

After training, block 25 transmits information about its results (including statistical data) to the control processor 28, which displays them on the displays of experts 32-34.

If the ANN training failed (information about this is displayed on the AWP of experts 32-34) or for some reason the experts are not satisfied with its results, then they can re-initiate the training process by giving the appropriate command to the control processor 28. At the same time, it is possible to reset everything synaptic weights of ANN 21 via interface 36 or set the required initial values for them.

After training, experts leave their jobs and do not participate in the further work of the system 20.

The warning system 20 in everyday operation works like this. The information receiving module 14 continuously receives via channels 12, 13 and leads to a uniform format and dimension information from sensors 4-11 installed on the slopes of mountains 2 and 3 as described above (see the block diagram in Fig. 6 )

The information receiving module 14 regularly supplies this information via a fiber optic cable 15 to the information inputs of a pre-trained (as described above) multilayer ANN 21 (located in the information transfer module 16), many of which correspond to many information channels for monitoring the current situation.

The final text message generation unit 22 monitors the status of the information outputs of ANN 21. Having not detected activity on them exceeding a predetermined threshold, unit 22 concludes that the emergency did not occur and does not perform any actions.

The warning system 20, in the mode of generation and transmission of short text messages to subscriber communication devices, operates as follows.

The final text message generation unit 22 detects activity at the information outputs of ANN 21 that exceeds a predetermined threshold, block 22 concludes that sensors 4-11 recorded something important, and ANN 21 received and classified their readings as related to an emergency, caused by mudflow.

Block 22 extracts from the bank of text segments 23 those text segments whose numbers correspond to the outputs of ANN 21, which are most activated.

From these text fragments, block 22 generates a single short text message, separating the text segments in the message using the same grammatical separator characters (in the preferred embodiment of the invention, the sign; “semicolon” is used for this) and making sure that the total text length is not exceeded the maximum permissible length for one short text message (restriction for SMS messages - no more than 70 characters for text in Russian and no more than 160 - for characters in Latin).

Schematically, this process is shown in the flowchart (see Fig. 7).

Information about the occurrence of an emergency and the generated text message proposed for transmission to the subscriber devices 19, the block 22 displays on the AWP 26 on-call duty MCC emergency department.

The operational duty officer from his workstation, if necessary, makes adjustments to the generated text message, determines that all areas of the city 1 are included in the area of probable emergency deployment (therefore, the message should be sent to all base stations located in the city), and confirms its consent to start the process of alerting residents of city 1 using system 20.

Schematically, this process is shown in the flowchart (see Fig. 8).

Having received consent from AWP 26 to start the notification process, block 22 transmits the generated text message via information channel 17 to the GSM network 18 deployed in city 1, indicating to the broadcast center (SHS) those base stations to which the message should be sent (to In our example, all the base stations of the city).

The same copies of this message are transmitted from the CBC from the names of the previously determined telephone number (or several numbers) belonging to the system 20, then to the BSC, which forwards it to the selected BTS base stations (see Fig. 5).

BTS broadcasts messages to all active subscriber communication devices 19 located in city 1, regardless of which group the subscribers belong to, of which “model” subscriber devices, and whether they are subscribers to the system or not.

The specified message is received by all active (i.e., included and located in the coverage area of the network 18) subscriber communication devices 19.

As an explanation, we give an example of a short text message generated by module 16 (the text is enclosed in quotation marks): “Attention !; Descent of small mudflows; From the eastern slope; Leave the lowlands. "

The author also has in mind the possibility of generating and sending SMS messages recorded in transliteration (a method of writing Russian text in Latin letters, as well as numbers and other available characters), which can increase the length of the message up to 160 characters (restriction for SMS messages - no more than 160 characters of the Latin alphabet).

The following is an example of a short transliterated text message generated by module 16 (the text is enclosed in quotation marks): “Vnimanie !; Shod malogo selya; S vostochnogo sklona; Pokinte niziny. "

Necessary explanations to the second aspect of the invention.

Block 45 tracking changes in the balance on the account of the mobile number of the subscriber caused by the exchange of messages between the subscriber unit and the warning system is connected to the information inputs / outputs of the system for receiving / sending messages of the subscriber device.

Block 45 is a programmable multi-function block, endowed with the ability to automatically (without user intervention) send and receive messages, transmit requests and commands to the mobile operator on behalf of the subscriber communications device, and also read the operator’s responses to them.

Block 45 has a memory in which it stores:

- IMEI number (International Mobile Equipment Identity) of the subscriber’s device, which requests the device during its initialization (for example, using the command * # 06 #);

- the telephone number of the subscriber in the GSM network (for example, 89094044387), which unit 45 requests from a mobile operator (for example, in Beeline this function can be implemented using the command * 110 * 10 #, for other mobile operators in a similar way) ;

- a short telephone number of the warning system, to which messages with codes should be sent to replenish the balance (for example, 7777).

Block 45 is equipped with a timer that periodically (every hour) prompts it to automatically request information from the mobile operator about the status of the account (balance) of the subscriber’s mobile number.

The timer can be implemented programmatically, by accessing procedures of the Time type, or by accessing the system clocks and time measurement services built into the subscriber device 19.

The last 45 answers of the operator (for example, four answers) to these requests block 45 stores in its memory (saves time and money in the subscriber’s account). The remaining answers are deleted from memory, new ones are written in their place.

For example, the function of a popular Russian operator Beeline can be realized by sending * 102 # requests and receiving corresponding answers from it. For other operators - in a similar way.

Block 45 is also configured to record the fact of receipt or sending messages to the address of the notification system from the subscriber unit (messaging). A flowchart of the operation of block 45 is shown in FIG. eleven.

If a message exchange with the notification system was recorded in the study period of time and the account balance decreased, then block 45 calculates the amount Sum, which should be replenished by the formula:

where Sum (n) is the balance at the current moment,

Sum (n-1) - balance recorded in the previous measurement.

Then block 45 generates a message with a unique confirmation code, which contains the following information:

- the amount of money to replenish the subscriber's account;

- number identifying the subscriber device;

- date and time of message generation.

The message format is standardized. An example message format is shown below: “IMEI; Subscriber number in the GSM network; Date of; Time; Sum. "

Block 45 places the specified message into the message sending / receiving system 51, which is built into the subscriber unit, and prompts him to send this message to a specific short number (for example, 7777) belonging to the notification system.

The system contains a module that automatically receives messages from subscriber communication devices containing unique confirmation codes to a certain short telephone number belonging to the system (for example, 7777).

The module in the available implementation is a server with software, a module for receiving / sending messages and an identification card for a GSM network (SIM) short number (for example, 7777), allowing it to receive, process, analyze SMS messages arriving at its address. And also endowed with the ability to send messages and command instructions to the operator (s) of communications operating in the city.

The module is configured to extract text from incoming SMS, check and analyze received confirmation codes. The indicated operation consists in the fact that the module checks the contents of the received messages for the presence in them and the correct location of all the required blocks. Messages that do not satisfy this condition are discarded and not processed.

For example, a message of the format “IMEI; Subscriber number in the GSM network; Date of; Time; Sum ”will be deemed true and will be processed. On the contrary, a message of the form “Sum; The subscriber number in the GSM network ”will be rejected (there are no IMEI, date and time, and the Sum block is not in its place).

Thus, the system 20 is protected from the actions of attackers trying to unreasonably replenish the account balance.

If the message is recognized as correct, the module extracts information about the required recharge amount and subscriber number from it, and then generates and sends an instruction to the operator company to transfer the specified amount from its account to the required subscriber number.

To implement this function, the module can use the Mobile Transfer service provided by Beeline operator.

This service allows you to transfer money from your account to the account of any other subscriber. To do this, the module must form an instruction command of the form: "* 145 * Recipient subscriber number in the GSM network * Transfer amount #".

You may need to confirm this operation by sending an SMS with confirmation to the operator's address. The module is configured to implement this operation in automatic mode.

It is also possible to implement this function using the "Mobile Payment" service from Beeline. The only difference is that this service allows you to replenish the account of any subscriber by transferring funds from the account of a registered bank card (VISA or MasterCard).

For this, the module must form an instruction command of the form: "* 100 * Secret code (issued upon registration) * Transfer amount * Number of the recipient in the GSM network #".

For other mobile operators, this function can be implemented in a similar way.

An approximate scheme of the module is shown in the block diagram (Fig. 12).

The balance of the account with the number 7777 belonging to the notification system is regularly (for example, monthly) replenished by the amount debited from it. The costs can be borne by local authorities (city budget), Ministry of Emergencies, sponsors, mobile operators, etc.

The operation of block 45 in a subscriber communication device is further illustrated in FIG. 9, and the operation of the entire system as a whole is illustrated in FIG. 10. Block 45 requests and maintains the balance of the subscriber’s account every hour by sending requests * 102 # and receiving corresponding answers.

In FIG. 9 shows that between 14:00 and 15:00 there was an exchange of messages with the warning system and the balance of the subscriber's account decreased. Block 45 starts the execution of the algorithm, the circuit of which is shown in FIG. 11. Block 45 calculates the sum Sum by the formula 5, which turns out to be equal to $ 3. This amount is “an amount equal to or slightly higher than debited from the subscriber’s account due to the exchange of information between subscriber devices and the warning system”.

Next, block 45 generates a message with a unique confirmation code of the following form: “IMEI; Subscriber number in the GSM network; Date of; Time; Sum. "

Block 45 places the specified message into the message sending / receiving system 51 (outgoing SMS queue 53) integrated in the subscriber unit and prompts him to send this message to the short number 7777 belonging to the notification system.

This message is received by the module, which automatically receives, to a certain short telephone number belonging to the system, messages from subscriber communication devices containing unique confirmation codes.

This module starts the execution of the algorithm, the circuit of which is shown in FIG. 12. If the message has the correct format, the module extracts the required amount ($ 3) and the subscriber number (89094044387) from the message text.

Further, the module generates an instruction team with a request to transfer funds from the account of the telephone number belonging to the notification system to the subscriber’s mobile number account using the Mobile Transfer service. The command has the following form: "* 145 * 9094044387 * 3 #".

The module sends the order to the telecom operator and, optionally, additionally sends a message or command confirming the money transfer operation.

This completes the operation of the module (account replenished), it monitors the receipt of the following messages from subscriber communication devices 19.

The author assumes that this function of system 20 should be protected from the actions of attackers trying to unreasonably replenish the balance of their account. To this end, the system may stop replenishing the accounts of those subscriber devices that exhibit unusual activity or are clearly abusing this function of the system 20.

Replenishment of the balance on the account of the mobile number of the subscriber by an amount equal to or slightly higher than written off due to the exchange of messages of the subscriber device 19 with the system 20 by sending to a specific short number a message containing a unique confirmation code is performed (as described previously), after the cessation of the development of the emergency situation and the restoration of normal operation of the wireless communication network of 18 city 1.

The author also assumes the possibility that block 45 will request a balance on the subscriber’s account (and save the results of the requests) not only regularly (for example, hourly), but every time it fixes the exchange of information between the subscriber unit and the notification system.

The more often balance requests will be made, the more accurately the required amount of replenishment will be determined, but this period of time should be chosen reasonably so as not to overload the GSM network and subscriber unit.

Necessary explanations to the third aspect of the invention.

Since the message sent by the system 20 contains text segments in a limited controlled version of the natural language, mainly used by residents of the geographic area to be notified, each of which carries information about a specific parameter of a possible emergency, the message’s meaning can be directly (without using additional technical means) perceived subscribers whose subscriber devices are not subscribed to system services for one reason or another.

Thus, subscribers who are not subscribers to the system, having received the corresponding text message from the telephone number of the notification service they know, will no longer consider it as a “meaningless line of code” (unlike the prototype system), but can easily understand its meaning without use any auxiliary means and begin to perform the necessary actions to save life / health / property.

At the same time, the subscriber communication device 19, owned by the subscriber of the notification system 20, provides, upon receipt of the same text message, the issuance of additional relevant information to the subscriber (in various forms - sound, graphic, text, etc.).

For this, the subscriber unit 19 must additionally contain the programmable multi-function blocks listed below (shown in Fig. 4), the logic of which can be implemented using hardware and / or software.

The data processing unit 38 is configured to check in the background all messages arriving at the SMS 51 sending / receiving system of the subscriber communication device 19. If the message did not come from the telephone number (or several numbers) of the warning system 20 (stored in block 38), then it does not performs no action.

If the message came from the number of the warning system, then block 38 suppresses the natural reaction of the subscriber device to the receipt of SMS and extracts the message from the incoming SMS queue 52 so that it is not processed by other systems of the device 19.

In the Android system, this function can be implemented using the abortBroadcast () operator.

Block 38 extracts the text of the message from SMS and processes it, dividing it into text segments. This function is implemented by sequential character-by-character viewing of the message text and the search for separator characters; ("semicolon"). The characters located between them are considered text segments, they are saved and numbered in the order from the beginning of the message to its end.

Block 38, using special tables and algorithms, translates each text segment into binary form (binarization of source data, i.e., translation of natural language words into a set of ones and zeros). This procedure is necessary because neural network 39 works with binary data.

Algorithms for converting text to binary are known and widely used. It is proposed to use the following method. The code of each text symbol in the ASCII table is calculated (for this, the standard function Ord () of the Pascal programming language can be used). Then, each decimal number received (character code) is converted to a binary number.

Thus, block 38 translates the text of the message, originally looking like “Text segment 1; Text segment 2; Text segment 3; Text segment 4; ... "in the form" Binary number 1; Binary number 2; Binary number 3; Binary number 4; ... " The conditional block diagram of the operation algorithm of block 38 is shown in FIG. 13.

Next, block 38 supplies the received binary numbers to the corresponding inputs of ANN 39, the set of which corresponds to the set of text segments (converted to binary numbers) contained in a short text message received from the notification system.

ANN 39 is made in the form of an ANN of the “multilayer perceptron” type with the number of hidden layers of at least two, which can be implemented both software and hardware (for example, as a dll module).

Many of the ANN outputs are numbered so that the output numbers correspond to the numbers of information components stored in the data bank 47. This point is illustrated in more detail in FIG. fourteen.

The data storage unit 47 is implemented as a data bank storing information components used to generate a notification displayed to the subscriber. Information components can be presented in various forms - sound, graphic, text, etc. For example, it could be:

- sound files with melodies to attract the attention of the subscriber or files containing the recorded voice of the speaker with the corresponding text;

- Graphic files with maps of the area, which show the places of evacuation and ways of approaching them;

- short videos with a clear demonstration of the use of protective equipment and the implementation of measures to save life / health / property;

- text files containing a verbal description of the relevant measures to save life / health / property, etc.

- profiles of vibration signals used to attract the attention of the subscriber;

- the frequency, color, and flashing brightness of the LED indicators and camera flashes built into the subscriber device (if available).

Each information component stored in block 47 has a unique number that corresponds to one ANN output number 39. In the simplest case, this correspondence can be established using a table similar to the one below.

This table is stored in block 46, while the information components themselves are in block 47.

The notification program generating unit 46 is located between the outputs of the ANN 39 and the data storage unit 47, configured to generate a notification to the subscriber from a plurality of information components corresponding to a plurality of activated information outputs of the ANN.

Thus, the main task of block 46 is to establish correspondence (according to table 3) between the numbers of the activated outputs of the ANN and the numbers of information components in block 47.

Having established the numbers of the corresponding information components, block 46 extracts them from block 47 and transfers them to block 48 for further processing. In this context, “notification program” should be understood as a set of information components transmitted by block 46 to block 48.

Block 48, prompting the subscriber device to perform certain actions corresponding to the generated subscriber notification program, is connected to the following elements built into the subscriber device: vibration alarm, speaker, LED indicator and flash of the built-in camera (if available), touch screen with on-screen keyboard.

Block 48, having received from block 46 a certain set of information components (“notification program”), is configured to immediately output them to the built-in means of the subscriber device, designed to process the corresponding types of information components:

- graphic and text files, videos - to screen 57;

- sound files - to the speaker 55;

- profiles of vibration signals - to the built-in vibration alarm 54;

- frequency, color and brightness of blinking - on LED indicators 56 built into the subscriber device and camera flash (if available).

The author suggests that block 48 should provide the subscriber with the ability to control the information components displayed on the screen using standard means of the subscriber device and its operating system. For example, minimize, close, maximize information windows, re-view (and listen) the corresponding notifications.

The number and composition of embedded sensors and subscriber unit systems (42-44) may vary depending on the subscriber unit model, as indicated in FIG. 4 using the ellipsis characters "...".

Blocks 42-44 can be, for example, a vibration sensor, a light sensor, a proximity sensor, a front camera, a microphone, positioning systems based on signals from base stations, radio emission measurement systems, temperature measurement systems, system clocks, etc.

In order to clarify and better understand the further provisions of the description, we give below the definitions of the following concepts.

Let us explain that under the "individual information about the subscriber" is understood, mainly, such information that remains unchanged for a relatively long period of time (months, years, decades) or does not change at all.

For example, this may include age (calculated by date of birth), the presence of significant permanent health problems (disabilities), the presence of special knowledge and skills, the presence of regular activities, habits and addictions (regular smoking, drinking alcohol), etc.

In contrast, “current information about the status of the subscriber and the environment” refers mainly to information that is most relevant at the time of the start of the notification and relatively quickly (minutes, hours) loses its relevance.

For example, the following points can be attributed to it: whether the subscriber is in a place of mass gathering of people (stadium, concert, etc.), whether the subscriber is currently experiencing short-term health problems (colds, high / low blood pressure), the subscriber is intoxicated, etc.

The unit for collecting individual information about the subscriber 50 is configured to display questions on the built-in screen of the subscriber device 57 and read user responses to them.

The survey of the subscriber in order to obtain "individual information about the subscriber" is made for the first time when the subscriber initializes the subscription to the services of the broadcast notification system. Then - several times a year (every 4-6 months) in order to establish possible changes in individual information about the subscriber (change in habits, the emergence of new knowledge and skills, the occurrence of disability, etc.).

Information can be read out by the subscriber’s answers to the relevant questions displayed on the touch screen 57, which can be formulated in various forms, for example: “Do you have a disability or other significant health problems? Yes / No ”,“ Do you work in the Ministry of Emergencies? Yes / No ”, etc.

The sets of questions and the diagrams of their output to the subscriber are stored in block 50. Questions are organized according to the principle of a “tree” of questions in such a way that positive or negative answers to some of them can cut off the following groups of questions (“branches”) and send the survey in a new direction.

For example, if the subscriber answers the question “Do you work in the Ministry of Emergencies? Yes / No ”, then he may be asked the following question“ Do you have special skills and knowledge in first aid? Yes / No, and then Do you have the right to evacuate people to asylum? Well no".

If the subscriber answers negatively to the question “Do you work in the Ministry of Emergencies? Yes / No ”, then the above clarifying questions are not displayed to him, but other groups (“ branches ”) of questions are displayed instead.

The number of questions displayed to the subscriber should be reasonable and sufficient. The author believes that this condition corresponds to a quantity not exceeding 100 questions.

At the same time, the user 50 does not save the user's answers to these questions, but in real time (almost immediately) generates them and transmits instructions to the ANS 40 training and configuration unit on how the ANS 39 should be configured and trained on the ANS 40.

This function is implemented by the fact that in block 50 a predefined set of logical rules of the form “If <set of conditions>, then <set of instructions to block 40 for training and configuring ANNs>” is stored.

Examples of such rules are given below. “If <Answer to the question of disability =“ YES ”and visual impairment =“ YES ”>, then <“ Zero the synaptic weights of the ANN outputs associated with the informational components of the graphic type ”and“ Increase the synaptic weights of the ANN outputs associated with the informational components sound type »>.

“If <The answer to the question about work in the Ministry of Emergencies =“ YES ”and Possesses the right to evacuate people to shelters =“ YES ”>, then <“ Increase the synaptic weights of the ANN outputs associated with information components that carry information about the places of evacuation of people ”>.

The unit for collecting current information about the status of the subscriber and the environment 41 is connected to the built-in sensors and systems of the subscriber device (42-44).

The unit for collecting current information about the status of the subscriber and the environment 41 regularly (in real time) collects current information about the status of the subscriber and the environment using the built-in sensors and subscriber unit systems (42-44).

Block 41 does not store the results of the collected data, but in real time (almost immediately) generates on its basis and transmits to the training and configuration block ANN 40 instructions on how block 40 should configure and train ANN 39.

The function is implemented (similar to the function of block 50) using a predefined set of logical rules (stored in block 41) of the form "If <set of conditions>, then <set of instructions to block 40 for training and configuring ANN>".

Examples of such rules are given below. “If <“ The built-in microphone detects a lot of extraneous noise ”and“ The light sensor detects the absence of bright light ”>, then <“ Decrease the synaptic weights of the ANN outputs associated with information components of sound type ”and“ Increase the synaptic weights of the ANN outputs associated with information components graphic type ">.

It also considers the possibility that block 41, if necessary, can display the subscriber on screen 57 clarifying issues related to his current condition and the environment recorded by sensors 42-44.

The rules of the form can answer for this: “If <“ The built-in microphone detects a lot of extraneous noise ”and“ The light sensor detects the absence of bright light ”>, then <“ Ask the question, is the subscriber in a place of crowded places? ”And“ If the answer = “Yes,” “Increase the synaptic weights of the ANN outputs associated with the information components responsible for evacuating from crowded places”>.

For example, if it detects characteristic noise, vibration, and a lot of radio emissions from other subscriber devices using the built-in microphone, unit 41 may ask the subscriber a clarifying question: “Now are you watching the match at the stadium? Well no". Examples of other similar questions are given below: “Are you doing your morning run now? Yes / No "," Are you currently experiencing health problems? Yes / No "," Did you drink alcohol today? Yes / No ”, etc.

The implementation of the functions of blocks 50 and 41 is generally similar. Both of them generate and transmit to the training and configuration unit ANN 40 instructions on how block 40 should configure and train ANN 39.

The differences are that block 50 gives such recommendations relatively rarely (once every 4-6 months, after a user’s survey), and block 41 very often (many times a day, every time sensors 42-44 record significant changes in the environment and subscriber status).

The ANS 40 training and configuration unit also does not store this information in itself, but real-time (almost immediately) records and stores the collected individual and current information inside the ANN by changing the ANN settings corresponding to the number of ANN layers, the structure of neuron connections and values synaptic scales (i.e., it actually trains and configures it) so that the ANN contained in the subscriber communication device activates a message from the notification system outputs corresponding to the set of information components most suitable for a particular subscriber and the environment, despite the fact that the content of messages sent by the notification system to all subscribers in the event of an emergency, is the same.

An alternative embodiment of a subscriber device is also considered, in which information about the subscriber's condition and environment is stored inside a separate ANN (or in several ANNs) located in a subscriber communication device and not related to the main ANN 39.

To implement the claimed function, block 40 must keep a table of correspondence between groups of numbers of information outputs of ANN 39 (and, consequently, numbers of information components stored in block 47) with groups that blocks 41 and 50 mention in their instructions about how block 40 should configure and train ANN 39.

For example, in block 40, information is stored that “ANN outputs associated with informational components of sound type” are outputs No. 2 and 3, and “ANN outputs associated with informational components responsible for evacuating from crowded places are outputs "No. 1, 4 and 7".

Referring to this table, block 40 changes the settings of the ANN corresponding to the number of layers of the ANN, the structure of neuron connections and the values of the synaptic weights (that is, it actually trains and configures it) so that the ANN 39 activates the outputs corresponding to the set upon receipt of a message from the warning system information components that are most suitable for a particular subscriber and environment.

Moreover, all the collected data about the subscriber and the environment is stored only in the subscriber device and is not forwarded anywhere, and the subscriber communication device, despite the fact that the contents of the messages sent by the system to all subscribers in case of emergency, is the same, notifies the subscriber when a message arrives from the notification system, in the most suitable way for a particular subscriber and a specific environment, and uses the necessary information components for this. That is, the reaction of each specific subscriber device belonging to a particular subscriber of the system to the receipt of such an alert will be individual.

It should also be mentioned that many users of previous similar systems refused to use them (or intentionally provided incorrect information about themselves) because of fears that their personal data might be transferred to third parties at the direction of certain intelligence agencies or stolen by attackers. This significantly slowed down the process of mass implementation of such systems.

This aspect of the invention is practically devoid of this drawback, not only because all personal data is stored only in the subscriber device and is not sent anywhere, but also because the very form of storing this data inside the ANN almost completely eliminates the possibility of its extraction, collection and systematization by attackers, even if the subscriber device is stolen.

The subscriber communication device 19, owned by the subscriber of the warning system 20, in everyday use, operates as follows.

The data processing unit 38 is connected to the SMS message receiving / sending system 51 of the subscriber terminal and, in the background, checks in the memory block 52 for storing incoming messages, all SMS messages received in the subscriber’s telephone to receive SMS messages from a specific number ( or several numbers) of the alert service 20.

Upon receipt of a short text message from the notification system informing about the occurrence of an emergency, the subscriber unit switches to the full functioning mode of all units and systems and alerts the subscriber about the occurrence of an emergency. For this, block 38 starts the execution of the previously described algorithm (block diagram in Fig. 13).

The training and configuration unit of the ANN 40 and the data storage unit 47 can receive (if necessary) the corresponding instructions from the notification system 20 and related services, instructing how to replenish and edit the data storage means 47 and train and configure the multi-layer ANN 39 stored in subscriber communication device.

Such instructions may include the recommended number of ANN layers, neural neural connections, and synaptic balance values. Instructions can be automatically requested and received from the update server when a subscriber’s device is connected to the mobile Internet, for example, via Wi-Fi.

The author also suggests the possibility of obtaining such instructions from short messages containing the necessary service information that the subscriber device 19 can request and receive from telephone numbers belonging to the warning system 20.

The unit 50 for collecting individual information about the subscriber asks the user in a dialogue mode for various information related to his personal data, at the first start of the subscriber device after subscribing to the services of the system 20, and then (if necessary) on a regular basis by outputting the appropriate requests to the touch screen 57 and reading user responses.

The requested data may include information on the date of birth (age is calculated on its basis), education, place of work, hobbies, hobbies, normal daily routines, sporting achievements, physical limitations (disabilities), place of permanent residence (including occupied floor) and etc.

Information can be read by the subscriber’s answers to the relevant questions displayed on the touch screen 57, which can be formulated in various forms, for example: “Do you work in the police? Yes / No ”, etc.

Unit 50 organizes and transmits the collected data to the ANN 40 training and configuration unit, which stores them inside the ANN 39 located in the subscriber communication device by changing the ANN settings corresponding to the number of ANN layers, the structure of neuron connections, and the values of synaptic weights, i.e. actually produces its training and configuration.

The unit for collecting current information about the status of the subscriber and the environment 41 regularly collects current information about the status of the subscriber and the environment using the built-in sensors and systems of the subscriber device (42-44), the number and composition of which varies depending on the model of the subscriber device, which is indicated on FIG. 4 using the ellipsis characters "...".

Unit 41 systematizes and transmits the results of the collected data to the ANN 40 training and configuration unit, which stores them inside the ANN 39 located in the subscriber communication device by changing the ANN settings corresponding to the number of ANN layers, the structure of neuron connections, and the values of synaptic weights, i.e. . actually produces its training and configuration.

It also considers the possibility that block 41, if necessary, can display the subscriber on screen 57 clarifying issues related to his current condition and the environment recorded by sensors 42-44.

The subscriber communication device 19, owned by the subscriber of the warning system 20, upon receipt of an emergency message from the warning system, operates as follows.

The system for receiving / sending SMS messages 51 of the subscriber unit receives the SMS message and stores it in the memory unit 52.

Block 38 checks the new message to see if it has been sent from a specific number (or several numbers) belonging to the notification system 20.

If not, no specific actions are performed, SMS is processed in the usual way, like any other message received from mobile subscribers, service providers, etc.

If the answer is yes (the message came from the notification system 20), then block 38 saves its text in its internal memory, deletes it from block 52, and suppresses the usual reaction of the subscriber unit to the receipt of any other SMS messages.

Block 38 divides the text of the message into text segments, denoted by a semicolon. For example, the message that was already cited above as an example of “Attention !; Descent of small mudflows; From the eastern slope; Leave the lowlands, block 38 will be divided into four text segments.

Block 38, using special tables and algorithms, translates each text segment into binary form (binarization of source data, i.e., translation of natural language words into a set of ones and zeros). This procedure is necessary because neural network 39 works with binary data.

Block 38 delivers the data thus obtained corresponding to the said text message segments to the information inputs of the multilayer ANN 39.

ANN 39 is trained by block 40 based on information received from 41 and 50 in such a way that it specifically responds to the received message.

The final generation unit of the subscriber alert program 46 reads signals from a plurality of activated ANN outputs and extracts those information components from the data storage means 47 that correspond to the activated ANN outputs.

At the same time, in block 47, which is a data bank, information components can be stored in various forms, for example, in graphic (graphic files with maps of the area on which the location of evacuation places is indicated), animated (videos describing various types of actions), audio ( sound files), digital (vibration profiles), tabular (addresses and phone numbers of the nearest medical institutions), etc.

Block 46, using these informational components from block 47, and using the information received from the information outputs of the ANN 39, forms a specific notification program that is most suitable for a particular subscriber and for a specific situation.

Having generated the notification program, block 46 transmits it to block 48, prompting elements of the subscriber device, such as a vibrator 54, speaker 55, LED indicators and camera flash 56, touch screen 57, to perform certain actions corresponding to the generated notification program.

For example, if the subscriber is in a brightly lit room (or is visually impaired), then to attract his attention it is inappropriate to use LED indicators 56, a flash of the built-in camera, as well as information displayed on the display 57. Instead, the focus will be on the most noticeable vibration signal for the vibrator 54 and an audio message for output to the speaker 55.

In this case, the author suggests that blocks 46, 47 and 48 can be tuned to synthesize human speech or reproduce an audio message composed of sound fragments stored in block 47.

Subscribers belonging to different target groups (police, military, doctors, rescuers, etc.) will also be notified by their subscriber devices 19 in a specific way.

It is also assumed that in response to receiving a regular text message (the same as for other subscribers), some subscriber devices 19 (for example, belonging to special services) have the ability to generate alerts that contain some secret data that is extracted (and / or generated) device systems from storage means 47 contained within the device. Such a notification can be issued to the subscriber taking into account the surrounding environment so as to prevent their leakage (for example, in a half-whisper or on a screen with dimmed brightness). When a device containing such data is stolen, it can automatically identify it and destroy it, for example, by completely resetting all synaptic scales in ANN 39 and cleaning block 47.

Subscriber devices can also be configured to notify the subscriber in his native language, even if he is in an area where another language is predominantly used.

For example, a subscriber who came to city 1 on an excursion from France can set up his subscriber device so that it will notify him with text and sound messages in French, despite the fact that SMS text messages sent by system 20 to all residents of city 1, formed in Russian.

The above examples are descriptions of only some of the possible options for responding to situations, many of which can arise in a real situation.

Necessary explanations to the fourth aspect of the invention.

Next, we describe how the means of preventing overload of the wireless communication network work, which, upon receipt of a short text message from the warning system 20, block in the subscriber communication device 19 the functions of making outgoing voice calls and sending messages.

The subscriber unit 19 further comprises a random number generator 49 connected to a block prompting the device to perform certain actions 48.

The embodiment of the patented device in block 49 uses thermal noise due to thermal motion of charge carriers in the conductor, which induces a fluctuating potential difference at the ends of the conductor, as a source of entropy.

In another implementation of the device, the random number generator can be performed by accessing the hardware source of entropy integrated in the control processor of the device 19.

The generation of random numbers by software is also possible, for example, using the standard functions Random () and Randomize of the Pascal programming language.

The block that prompts the device to perform certain actions 48 in addition to the functions that were described for him earlier, has additional functionality.

Namely, block 48 is configured to block the function of making outgoing voice calls and sending messages for a period of time from 3 to 15 minutes from the moment of receiving a text message, which is randomly selected in the specified range, using the short message from the notification system to block the subscriber’s device random number generator.

And also block 48 is configured to display a notification on the screen and speaker of the subscriber device informing about the impossibility of outgoing voice calls and sending messages, as well as about the time remaining until the end of the block.

A block diagram of the operation of block 48 is conventionally shown in FIG. 15. The signal to start this algorithm for block 48 is the receipt of informational components from block 46. After that, the block calls the generator 49 and receives a real number from it (the range of possible numbers is limited to 3-15).

Block 46 converts the number into a time period of the form C minutes: D * 60 seconds, where C is the integer part of the generated number, and D is the fractional part of the number. For example, the number 4,5 corresponds to a time period of 4 minutes 30 seconds, and the number 6.2 corresponds to a time period of 6 minutes 12 seconds.

Block 48 blocks in the subscriber communication device the functions of making outgoing voice calls and sending messages for the generated time period as follows. To do this, he deactivates on the touch screen 57 and the built-in on-screen keyboard portions of the screen with keys responsible for making outgoing voice calls and sending text messages. These sections of the screen can be closed by messages pop-up on top of them (function windows) or displayed in a dull (gray) color.

Block 48 contains a timer or counter that updates (decreases) the time remaining until the end of the blocking every minute, and in the last minutes of blocking every ten seconds. The timer can be implemented programmatically, by accessing procedures of the Time type, or by accessing the system clocks and time measurement services built into the subscriber device 19.

Block 48 is configured to display on the screen and speaker of the subscriber unit a notification prepared in advance and stored in it, informing about the impossibility of outgoing voice calls and sending messages, as well as about the time remaining until the end of the blocking.

The indicated period of time is updated in this message by a timer (counter) command. This message should be displayed so as not to interfere with the subscriber to perceive other information components displayed on the screen of the device 19 in order to encourage the subscriber to take actions to save life / health / property.

For example, a message in the notification area of the Android operating system corresponds to this condition.

At the end of the blocking time (calculated using the timer), this message is deleted from the screen, and the device’s functions for making outgoing voice calls and sending messages are restored in full.

Next, we give an example of the operation of this function of the device 19.

Block 48 immediately before the start of the specific notification program received from block 46, refers to the random number generator 49, which can be implemented in hardware or software. The random number generator 49 generates a real number from the range [3; fifteen]. For example, it could be the number 7.5.

Having received this number, block 48 blocks in the subscriber communication device 19 the functions of making outgoing voice calls and sending messages for a period of time corresponding to the real number generated in 49. For our example, this will be 7 minutes 30 seconds.

Having performed this action, block 48 immediately starts executing a specific notification program received from block 46, and, in addition, displays to the subscriber information about the impossibility of outgoing voice calls and sending messages, as well as about the time remaining until the end of the block.

For example, such a message may look like this: “The functions of making outgoing calls and sending messages are blocked. 7 minutes 30 seconds are left until the end of the lock. Do not try to reach relatives and friends yet. Take it easy - they will be notified. Follow the instructions. "

The sentence “Until the end of the lock is left ...” will be updated on the screen 57 of the subscriber device 19 every minute, and at the last minute of the lock - every ten seconds, informing about the remaining time until the end of the lock.

Returning to the example of city 1, where the patented system 20 is deployed, and about 30% of the inhabitants are its subscribers, the technical solution described above will reduce the “call wall” in the first 3-15 minutes after an emergency by about 6-7 thousand calls. This will save the wireless network 18 from overload and failure, will allow timely delivery of all messages to the recipients.

Due to the generation of blocking time using a random number generator, communication devices will not return to the network all at once, but gradually. Moreover, according to studies, by this time most of the warning messages will already be delivered and the panic among subscribers should decline.

At the same time, users of subscriber devices who are not subscribers of the notification system 20 will be able to make outgoing calls and send SMS in the usual way.

After the end of the time period corresponding to the real number generated by block 49, the subscriber unit 19 restores the ability to make outgoing calls and send messages, about which the user is notified accordingly.

Having completed the necessary notification program, block 48 is turned off, and the subscriber unit restores its normal operation. It is also possible to prematurely terminate the notification program when the subscriber device 19 receives the corresponding short message “End !; The danger has passed ”from the notification system number 20, or stopping the notification program at the initiative of the user by pressing a special combination of keys on the touch screen with an on-screen keyboard 57.

Necessary explanations to the fifth and sixth aspects of the invention.

In these aspects of the invention, a Method for the operation of a broadcasting system for alerting mobile subscribers about emergencies in the mode of generation and transmission of short text messages to subscriber communication devices, as well as A method for operating a subscriber communication device belonging to a subscriber of a broadcasting system for alerting mobile subscribers about emergency situations is described upon receipt of an emergency message from the warning system.

The features of these aspects of the invention, in fact, are a combination of the functions of the system and subscriber devices included in it, described in detail earlier in the aspects of the invention from first to fourth.

Therefore, all of the explanations made by the author for the first to fourth aspects of the invention are fully true with respect to the corresponding features of the fifth and sixth aspects of the invention.

Necessary explanations for the seventh aspect of the invention.

Another significant function of the device 19 is that it can additionally carry out the following actions: register a clear difference between the actions actually performed by the subscriber and the recommendations issued by the communication device 19 upon receipt of a message from the notification system; display a request to the subscriber whether he considers the actions actually performed by him to be correct; record actions actually performed by the subscriber; to make a subscriber evaluate the correctness of the actually performed actions; to train and configure the ANN 39 stored in the subscriber communication device based on the actions actually recorded by the subscriber taking into account the assessment of their correctness in order to eliminate the differences between the recommendations issued and the actions actually performed by the subscriber in case of identical emergency situations in the future.

This allows you to gradually improve the operation of the device 19 due to the most important property of the ANN 39 stored in it - the possibility of learning by examples.

These device functions can be realized through the use of an additional software application stored in the memory of the subscriber device 19 (for example, a program for the Android operating system) or through another programmable unit embedded in the subscriber device 19.

The specified block or program should be able to directly access blocks 38-57 shown in FIG. 4. It is assumed that all of the functions described below of the subscriber device 19 implements this unit or program.

The function of registering a clear difference between the actions actually performed by the subscriber and the recommendations issued is carried out as follows.

The device 19 stores sets of signs (in the form of logical rules) characterizing typical signs of a subscriber's performance of various recommendations.

For example, if the subscriber is recommended to evacuate to shelter, then this action is characterized by such signs as are recorded by sensors 42-44, such as:

- increased vibration when running;

- the presence of extraneous noise.

Thus, the rule may look like: "If <vibration is greater than some value> and <extraneous noise is greater than some value>, then <subscriber is evacuated>, otherwise <subscriber is NOT evacuated>".

Suppose a device 19 received a message from a public address system, and it responded to it as described in the third aspect of the invention. Certain outputs of ANN 39 were activated and the corresponding information components from block 47 were issued to the subscriber.

These outputs (and information components) are responsible for advising the subscriber to evacuate to cover. This fact can be established due to the fact that in block 40 information is stored (in tabular form) that "the ANN outputs associated with the information components responsible for the evacuation are outputs No. 1, 4 and 7" (see explanations to the third aspect).

The device 19 refers to the readings of the vibration sensors and the built-in microphone 42-44, comparing them with the above rule. The device determines that the conditions <vibration greater than a certain value> and <extraneous noise greater than a certain value> are not satisfied. Based on this, it is concluded that <the subscriber is NOT evacuated> (the rule branch after the word "otherwise ...").

Thus, the function of registering a clear difference between the actions actually performed by the subscriber and the recommendations issued is considered to be implemented.

In this case, the device 19 immediately displays on the screen 57 a request message “Are you recommended to evacuate, but you are not following the recommendations? Well no". The subscriber answers “Yes.” The next question is “Do you think the actions you are actually doing are correct? Well no?". The subscriber answers “Yes.” The next question is “Why are you doing this?” The caller replies, “My legs hurt. It will be better if I stay in the building. ”

Then, in the subscriber device 19, the function “fixing the actions actually performed by the subscriber” is activated.

To do this, a white field with a blinking cursor and an on-screen keyboard are displayed on screen 57. The subscriber is invited to consistently (on points: 1, 2, 3, 4 ...) describe in a text all the actions performed by him.

For example, the subscriber writes: “1. I drew the curtains; 2. Turned off the gas and electricity; 3. He took the medicine; 4. Hid under the table. " The corresponding text file is generated.

At the same time, the built-in microphone can be activated, and the subscriber is invited to say aloud (dictate, describe) all the actions he performs. Corresponding audio files are generated.

At the same time, the built-in camera can be activated, and the subscriber is invited to visually show (take a photo and video) all the actions he performs. Corresponding photo and video files (graphics, video) are generated.

Thus, the function of fixing the actions actually performed by the subscriber is considered implemented.

After some time, for example, after receiving the signal “End !;; The danger has passed ”, the function of the subscriber device 19“ assessment by the subscriber of the correctness of the actually performed actions ”is activated.

For this, files created earlier by the subscriber with a description of the actions performed by him in case of an emergency are displayed on screen 57 and speaker 55. The subscriber is asked: “Do you think the actions you performed are correct? Well no".

If the subscriber considers the actions performed by him erroneous, then the subscriber can make adjustments to the existing files with recommendations. For example, add the item: “5. Do not forget to take the documents! "

If the subscriber considers the actions performed by him to be correct, then these files with a description of the actions actually performed by him are recorded in block 47, becoming the new information components stored in block 47.

At this function, the subscriber evaluates the correctness of the actions actually performed by him is considered implemented.

Then, in the subscriber device 19, the function “training and configuration of the ANN stored in the subscriber communication device based on the actions actually recorded by the subscriber taking into account the assessment of their correctness” is activated.

For its implementation, the ANS 40 training and configuration unit provides control instructions to change the settings of ANN 39 (corresponding to the number of layers of the ANN, the structure of neuron connections and the values of synaptic weights) so that upon receipt of an identical message from the notification system, ANN 39 activates the outputs corresponding to many new information components in block 47 created in the previous step.

To this end, block 40 can sequentially supply the same message (its text segments converted to binary) to the input of ANN 39, which it received earlier when the emergency was investigated, and change the ANN settings until it achieves so that the ANN 39 outputs are activated corresponding to the set of new information components in block 47 created in the previous step.

Thus, there is a “elimination of differences between the recommendations issued and the actions actually performed by the subscriber in the event of identical emergency situations in the future."

At the same time, the author understands that actions related to finding out the reasons for certain actions of the subscriber should not prevent the subscriber from taking the necessary actions to save life / health / property during the development of an emergency. The actions described above should be carried out so as not to interfere with the subscriber unit to perform other necessary actions during the development of an emergency.

It also implies the possibility of fixing the actions actually performed by the subscriber, assessing their correctness, and training the ANN when training alarms are received. In this case, training alarms periodically conducted by the relevant services will contribute to self-learning and increase the effectiveness of actions not only of the subscriber, but also of his subscriber device.

It is also possible to form special groups of people, based on the actions of which, in the manner described above, reference information components are generated upon receipt of various training alarm signals, which are integrated into all other subscriber devices of 19 subscribers of the warning system. Thus, the formation of information components stored in block 47 can occur.

Final provisions.

The author of the invention provides for the possibility that the subscriber device can store in its memory the previously generated action programs that were not completed for one reason or another and display them later at the initiative of the user (the function of viewing and analyzing outdated alerts).

The author hopes that as a result of the development of computational capabilities, ANNs will be able to automatically perform semantic analysis of texts compiled not in a controlled version, but in a full-fledged natural language.

Moreover, the author suggests that the subscriber device may have elements of the functionality of such a device as a lie detector (polygraph) - a technical tool that records changes in the physiological parameters of the person being interviewed to assess the reliability of the information communicated by him.

Registration of changes in the physiological parameters of the subscriber during a survey conducted by block 50 can be discreetly performed using built-in sensors and subscriber unit systems (42, 43, 44) that record, for example, shake of the subscriber’s hands (using a vibration sensor), change in facial expression ( using the front camera), slurred speech (using a microphone), etc.

The sudden appearance of such signs in the behavior of the subscriber may indicate an attempt by the subscriber to give the device a deliberately false answer.

The logic of the reaction to the appearance of such signs in the behavior of the subscriber when answering individual questions can be embedded in the subscriber device 19 using logical rules of the form "If <a set of conditions>, then <recognize the subscriber’s answer as not trustworthy>".

If the response of the subscriber is deemed not trustworthy, then the subscriber device 19 can invert the response of the subscriber. That is, instead of the answer “yes”, consider that the subscriber answered “no” and vice versa.

Thus, it becomes not so important whether the subscriber answers a specific question “yes” or “no”. Unit 50 of the subscriber unit, according to the response of the interviewee, will be able to independently determine the most reliable answer to the question.

The author also considers the possibility that during the polling of the subscriber, a three-dimensional virtual assistant image will be displayed on the screen 57, identified by the subscriber with himself, a close relative, a popular character or pet, displayed as an animated image of a human or animal figure.

The virtual assistant can present various “incentives” presented during the survey, or simulate the activity and emotional state of the person asking the questions. For example, a virtual assistant can tell a subscriber who identifies him with a relative the following phrase: “Anya, tell mom only the truth ...”.

Moreover, the author suggests that block 41 may also have elements of the functionality of a “lie detector” (polygraph), as described above for the functions associated with block 50.

It also assumes the possibility of recording changes in the physiological parameters of the subscriber during the implementation of the recommended actions in case of emergency, and issuing additional instructions to him related to his emotional and physical condition.

For example, if a subscriber is overly scared and makes rash actions (a vibration sensor detects hand tremors, a front camera detects pupil expansion, a microphone detects loud jerky speech), then an additional message “Calm down, everything will be fine, may be displayed on screen 57” Follow the instructions, ”and soothing music is displayed on speaker 55.

It is also possible to issue recommendations for taking medications based on the state of the subscriber and the environment, for example: “Remember to take a hypotensive drug” or “Refrain from sleeping pills.”

Implementation of the main provisions of the patented system is possible in the near future. At the moment, the author is developing a prototype of the described implementation of the invention, which includes subscriber devices running the Android operating system. Significant results have been achieved in the training of various ANNs.

It is understood that many further changes may be made by those skilled in the art regarding the details of the invention that have been described and illustrated above to explain the invention, without going beyond the scope and scope of the present invention as set forth in the description and claims.

The applicant hopes that the implementation and implementation of the invention will help in the noble cause of saving people from death and personal injury as a result of the sudden occurrence and development of emergency situations. For his part, the applicant is ready to make every possible effort.

BIBLIOGRAPHIC DATA OF INFORMATION SOURCES

1. Cell Broadcast (CB - Broadcast messaging) [Electronic resource]. URL: http://celnet.ru/CB.php

2. Information and communication technologies for life safety: monograph / under total. ed. P.A. Popov, EMERCOM of Russia. - M .: FSI VNII GOCHS (FC), 2009 .-- 272 p.

3. Security [Electronic resource]. URL: http://www.mgovservice.ru/industry/security.php (accessed date: 09/05/2013);

4. Where there is smoke, there is fire [Electronic resource]. URL: http://www.lesovod.org.ua/node/13667 (accessed: December 29, 2013).

5. Makarychev M. Cellular defense // Russian newspaper. Federal issue. - No. 5475 (99) (http://www.rg.ru/2011/05/11/znaki-site.html).

6. Fathutdinov R. Homeland hears, Homeland knows. In the Republic of Tatarstan, a fundamentally new population warning system has been launched // Bulletin of the Russian Ministry for Emergencies. - Moscow, 2013. - No. 3 (61). - S. 22-25.

7. Tsurikov A.N. Improving the technology of targeted emergency notification using SMS messages // Scientific and Technical Bulletin of the Volga Region. - Kazan, 2013. - No. 1. - S. 287-291.

8. Wienand Neil Heinrich Information broadcasting system // US Patent, US 2010/0202337 A1, published 08/12/2010.

9. Winand Neil Henry Information Broadcasting System // RF Patent, RU 2501091 C2, published December 10, 2013.

10. Kang D.I. The method of providing financial transaction services using a mobile phone // Eurasian patent No. 013142, published 02.26.2010, Bulletin No. 1.

11. Beeline subscriber directory (prepaid billing system).

Claims (7)

1. Broadcast warning system for mobile subscribers about emergencies, including a set of tracking tools for the current situation; an information receiving module designed to receive information from means for monitoring the current situation; an information transmission module for receiving information from the information receiving module and generating short messages that contain encoded data for transmission to subscriber communication devices in at least one pre-selected geographical area; subscriber communication devices belonging to subscribers of the system, receiving messages with encoded data and generating notifications for the subscriber corresponding to at least one information component stored in said subscriber communication devices, characterized in that
the system owns a single set of telephone numbers registered and supported by all local, regional and global mobile operators operating in the territory of the system’s deployment;
the information receiving module is configured to bring the information received from the means of tracking the current situation to a single format and dimension;
the information transfer module is configured to transmit absolutely identical copies of the message generated by it to all subscriber communication devices located in the geographical area of emergency deployment, regardless of which group the subscribers belong to and whether they are subscribers to the system or not;
the information transfer module contains a data bank of text segments in a limited controlled version of the natural language, mainly used by residents of the geographical area to be notified, each stored text segment carries information about a specific parameter of a possible emergency situation, and,
a trained multilayer artificial neural network (ANN), the set of information inputs of which corresponds to the set of information channels of tracking tools for the current situation, and the set of outputs corresponds to the set of text segments stored in the data bank of text segments used to generate short text messages intended for transmission to subscriber communication devices and also contains
the final text message generation unit located between the ANN outputs and the text segment data bank, configured to generate a message intended for transmission to subscriber communication devices by
extracting from the data bank those text segments that correspond to the activated ANN outputs,
separation of text segments using the same grammar separator characters,
restrictions on the length of the text of the generated message so that it does not exceed the maximum permissible length for one short text message;
means ensuring the coordination of the beginning of the notification process with the authorized body of authority and an indication of the geographical area of the probable deployment of an emergency to be notified,
connected to the final text message generation unit;
and also contains tools for training the ANN and its configuration, including:
control processor with a set of commands and custom software,
random number generator,
generator of visual and audio images,
databank of images and sounds,
memory block
means for displaying and inputting information with the ability to read expert opinions and store weighting factors assigned to experts,
block for calculating the majority of votes of the interviewed experts, taking into account the assigned weighting factors,
block correction of synaptic weights of ANN neurons in accordance with a given correction step,
compatible interfaces for connecting to ANN,
at the same time, the functional relationships between the above training and configuration tools of the ANN are organized so that
information inputs / outputs of the control processor are connected to the corresponding outputs / inputs of a random number generator, a visual and audio image generator, a memory block, a synaptic weights correction unit for ANN neurons, a majority vote block of interviewed experts, a text segment data bank,
wherein the information inputs / outputs of the majority vote block of the interviewed experts are connected to the outputs / inputs of the information display and input devices,
also the information output / input of the image and sound data bank is connected to the input / output of the generator of visual and audio images,
and the information output of the memory block is connected to the input of the block of correction of the synaptic weights of the ANN neurons,
the indicated block for the correction of the synaptic weights of the ANN neurons is also connected to the following interfaces for connecting to the trained neural network:
an output for supplying training input signals to the neural network,
output for correction of the vector of synaptic weights of the neurons of the trained neural network and
an input for reading the output signals of a trained neural network. 2. Broadcast warning system for mobile subscribers about emergencies according to claim 1, characterized in that
subscriber communication devices belonging to subscribers of the system further comprise a unit for tracking changes in the balance on the account of the subscriber’s mobile number caused by the exchange of messages between the subscriber device and the notification system,
information inputs / outputs of which are connected to the system for receiving / sending messages of a subscriber device,
the specified unit is configured to periodically automatically request from the mobile operator of the subscriber information on the status of the account of the mobile number of the subscriber and,
by registering the debit from the account caused by the exchange of messages between the subscriber unit and the notification system,
generates a message with a unique confirmation code, which contains at least the following information:
the amount of money to be replenished by the subscriber’s account,
subscriber unit identification number
the date and time the message was generated, and
prompts the device to send this message to a specific short number belonging to the notification system,
and also, characterized in that the system comprises a module that automatically receives, to a specific short telephone number belonging to the system, messages from subscriber communication devices containing unique confirmation codes,
verification and analysis of these confirmation codes,
and replenishment of accounts of mobile numbers of subscribers for an amount equal to or slightly higher than debited from their accounts due to the exchange of information between subscriber devices and the warning system,
by transferring funds from the account of the telephone number belonging to the notification system to the account of the subscriber’s mobile number. 3. A subscriber communication device belonging to a subscriber of a broadcast system for alerting mobile subscribers about emergencies, comprising a data storage unit in which a plurality of information components are stored; data processing means that are configured to receive short messages from the warning system, characterized in that the subscriber device contains:
a data processing unit configured to receive short text messages that extracts text segments from a message received from telephone numbers of the warning system, in a limited controlled version of the natural language, each of which carries information about a specific emergency parameter, and is indicated by the same grammatical characters- delimiters, and converting each text segment to binary;
a multilayer artificial neural network (ANN), the set of information inputs of which corresponds to the set of text segments contained in short text messages coming from telephone numbers of the warning system, and the set of outputs corresponds to the set of information components contained in the data storage unit;
a unit for collecting individual information about the subscriber, configured to display questions on the built-in screen of the subscriber device and read user responses;
a unit for collecting current information about the status of the subscriber and the environment using the built-in sensors and systems of the subscriber device;
an ANN training and configuration unit configured to record and store collected individual and current information inside the ANN by changing the ANN settings corresponding to the number of ANN layers, the structure of neuron connections, and the values of synaptic weights in this way
so that the ANN contained in the subscriber communication device activates the outputs corresponding to the set of information components most suitable for a particular subscriber and the environment upon receipt of a message from the notification system, despite the fact that the contents of the messages sent by the notification system to all subscribers in case of emergency equally;
a notification program generating unit located between the ANN outputs and the data storage unit, configured to generate a notification to the subscriber from a plurality of information components corresponding to a plurality of activated ANN information outputs;
a unit inducing the subscriber device to perform certain actions corresponding to the generated subscriber alert program;
while the functional connections between the elements of the subscriber communication device are organized so that the input of the data processing unit is connected to the system for receiving / sending messages of the subscriber device, and the outputs are connected to the information inputs of the ANN,
the ANN training and configuration unit is connected to the following interfaces connected to the ANN:
an output for supplying input signals to the ANN,
output for changing ANN settings and
an input for reading the output signals of the ANN,
a unit for training and configuring the ANN is also connected to a unit for collecting individual information about the subscriber and a unit for collecting current information about the status of the subscriber and the environment,
the unit for collecting individual information about the subscriber is also connected to the built-in screen of the subscriber device and the keyboard,
a unit for collecting current information about the status of the subscriber and the environment is connected to the built-in sensors and systems of the subscriber device;
the notification program generation unit is connected to the outputs of the ANN and the data storage unit, the output of the notification program generation unit is connected to the input of the unit that prompts the subscriber device to perform certain actions,
the unit that prompts the subscriber device to perform certain actions is connected to the following elements built into the subscriber device:
vibration alarm, speaker, LED indicators and flash of the built-in camera, screen with keyboard. 4. A subscriber communication device owned by a subscriber of a broadcast system for alerting mobile subscribers about emergencies according to claim 3, characterized in that it further comprises
a random number generator connected to a unit causing the device to perform certain actions, and
the unit that prompts the device to perform certain actions is configured to block, in the subscriber communication device, the functions of making outgoing voice calls and sending messages for a period of 3 to 15 minutes from the moment of receiving the corresponding text message, which is randomly selected when receiving a short text message from the notification system in the specified range using a random number generator, and
configured to display on the screen and speaker of the subscriber device a notification informing about the impossibility of outgoing voice calls and sending messages, as well as about the time remaining until the end of the blocking. 5. The method of functioning of the broadcast system for notifying mobile subscribers of emergency situations in the mode of generation and transmission of short text messages to subscriber communication devices, which includes the following steps:
register and maintain a single set of telephone numbers belonging to the warning system, with all local, regional and global mobile operators operating in the territory of the deployment of the system;
receive and lead to a single format and dimension information from means of tracking the current situation;
provide information from the means of monitoring the current situation to the information inputs of a pre-trained multilayer artificial neural network (ANN), many of which correspond to many information channels of means for monitoring the current situation;
using INS, generate a single short text message in a limited controlled version of the natural language, mainly used by residents of the geographic area to be notified, from text segments, many of which are stored in the text segment data bank and correspond to many ANN outputs, moreover
the message text is divided into a certain number of segments using the same grammatical separator characters, and each segment carries information about a specific emergency setting, and
the total length of the text is chosen so that it does not exceed the maximum permissible length for one short text message;
determine the geographical area of the likely deployment of an emergency;
coordinate the beginning of the notification process with the authorized authority;
transmit identical copies of a short text message previously generated by the ANN to all subscriber communication devices located in a pre-selected geographical area of emergency deployment, regardless of which group the subscribers belong to and whether they are subscribers to the system or not;
receive these messages to subscriber communication devices;
receive messages from subscriber communication devices containing unique confirmation codes to a specific short telephone number;
read and verify the specified confirmation codes;
replenish accounts of mobile numbers of subscribers for an amount equal to or slightly higher than debited from their accounts due to the exchange of information between subscriber devices and the warning system,
by transferring funds from the account of a telephone number belonging to the notification system to the accounts of mobile numbers of subscribers. 6. The method of functioning of a subscriber communication device belonging to a subscriber of a broadcasting system for alerting mobile subscribers about emergencies when a message about an emergency occurs from a warning system, which includes the following steps:
receive a short text message from the notification system;
extracting text segments from the message, each of which carries information about a specific emergency parameter, and is indicated using the same grammatical separator characters;
translate each text segment into binary;
serving the selected text segments to the information inputs of a multilayer artificial neural network (ANN) corresponding to the text segments of the message;
reading signals from the multiple outputs of the ANN corresponding to the plurality of information components;
extracting those information components from the data storage means that correspond to the activated information outputs of the ANN;
start the execution of certain actions of the subscriber device to alert the subscriber corresponding to the extracted information components and the activated information outputs of the ANN;
randomly select a period of time from a range of 3-15 minutes using a random number generator;
block in the subscriber communication device the functions of making outgoing voice calls and sending messages for a period of time corresponding to the period selected in the previous step;
display on the screen and speaker of the subscriber device information about the impossibility of outgoing voice calls and sending messages, as well as about the time remaining until the end of the blocking;
restore normal operation of the subscriber device in everyday use mode;
monitor changes in the balance on the account of the mobile number of the subscriber caused by the exchange of messages between the subscriber device and the notification system, and
replenish it by an amount equal to or slightly higher than written off, by the fact that
generate and send to a specific short number a message with a unique confirmation code containing at least the following information:
the amount of money to be replenished by the subscriber’s account,
subscriber unit identification number
Date and time the message was generated. 7. The method of operation of a subscriber communication device owned by a subscriber of a broadcast system for alerting mobile subscribers about emergencies when a message about an emergency occurs from the notification system according to claim 6, characterized in that it further includes the following steps:
registering a clear difference between the actions actually performed by the subscriber and the recommendations issued by the communication device when a message is received from the notification system;
display a request to the subscriber whether he considers the actions actually performed by him to be correct;
fix the actions actually performed by the subscriber;
the subscriber evaluates the correctness of the actions actually performed by him;
they train and configure the ANN stored in the subscriber communication device based on the actions recorded by the subscriber actually performed taking into account the assessment of their correctness with the aim of
eliminating the differences between the recommendations issued and the actions actually performed by the subscriber in case of identical emergency situations in the future.

Images